DESCRIPTION
[Title of Invention]
CODING DEVICE, DECODING DEVICE, CODING METHOD,
DECODING METHOD, AND COMMUNICATION SYSTEM
[Technical Field]
[0001]
The present invention relates to coding performed with use of quasi-cyclic
codes and corresponding decoding, and in particular to technology for preventing an
erroneous code word from being judged as error-free. Such a wrong judgment is
herein referred to as a false detection.
[Background Art]
[0002]
In communication systems such as a wireless LAN (Local Area Network)
and PLC (Power Line Communication), when a transmitter device transmits data to
a receiver device, an error occurs in the data received by the receiver device at a rate
depending on the status of a channel. There are countermeasures against such an
error that could occur in data upon transmission/reception of the data. One example
of such countermeasures is a retransmission control method. According to this
method, the receiver device checks whether the data it has received contains an error,
and if the data contains an error, the data is retransmitted from the transmitter device.
This way, error-free data will eventually be transmitted from the transmitter device
to the receiver device.
[0003]
This retransmission control method is described blow by taking an example
of communication conforming to the communication standard IEEE (the Institute of
Electrical and Electronics Engineers) 802.1 In. FIG. 25 shows a PHY frame structure
according to IEEE 802.1 In. In a PHY frame shown in FIG. 25, a preamble signal
composed of the following items is appended to the front of data to be transmitted
(i.e., data carried by a DATA field in FIG. 25): an 8-u.s L-STS (Legacy-Short
Training Symbol); an 8-fj.s L-LTS (Legacy-Long Training Symbol); a 4-us SIG
(SIGnal field); an 8-us H-SIG (High throughput-SIGnal field); a 4-us HSTF (High
throughput Short Training Field); and a 4-us HLTF (High throughput Long Training
Field) 1.
[0004]
A preamble signal is used by the receiver device to perform time and
frequency synchronization, AGC (Automatic Gain Control), and channel estimation.
As a preamble signal contains information that is required to demodulate the DATA
field, such as the length of data to be transmitted and the modulation/coding method
for the data, the preamble signal is always appended at the front of a PHY frame.
Furthermore, an error detection code called a CRC (Cyclic Redundancy Check) code
is affixed to the data to be transmitted via the DATA field.
[0005]
After the receiver device receives a PHY frame and performs predetermined
signal processing, it transfers the received data to a MAC layer processing unit that
performs predetermined processing in a MAC layer. The processing performed in
the MAC layer includes an error detection using the CRC code appended to the
received data. If the received data is judged as error-free as a result of the error
detection, then the receiver device transmits, to the transmitter device, an
acknowledgement response (ACK: ACKnowledge) indicating an error-free data
reception. On the other hand, if the received data is judged as containing an error as
a result of the error detection, then the receiver device does not transmit the ACK.
When the transmitter device has not received the ACK corresponding to the PHY
frame for a predetermined time period since the completion of transmission of the
PHY frame, it judges that the data has not been properly received by the receiver
device and accordingly retransmits the same data. Through the above procedures,
the transmitter device repeatedly transmits the same data until the receiver device
receives the data with no error. As a result, reliability of data transfer from the
transmitter device to the receiver device can be improved.
[0006]
Note that the retransmission control can also be achieved when the receiver
device is configured to transmit a retransmission request, which is a signal for
requesting retransmission of a data frame upon detection of an error, to the
transmitter device. In this case, the transmitter device retransmits the data frame
specified by the retransmission request upon receiving the retransmission request.
[0007]
However, when a PHY frame is long, i.e., when the data size of a DATA
field is large, the above-described error detection and retransmission control that are
performed on a per-PHY frame basis are problematic as they cause reduction in the
throughput. This is because the transmitter device would need to retransmit the long
PHY frame in its entirety even when only part of the data received by the receiver
device contains an error due to a local fluctuation of a wireless communication
channel. Moreover, the longer a PHY frame is, the longer it takes to perform a single
retransmission. Given this fact, when there are a plurality of terminals operating as
transmitter devices or receiver devices, the throughput of the entire system is
reduced.
[0008]
As one method to solve this problem, a subblock retransmission method has
been studied. The subblock retransmission method divides a DATA field into small
blocks and performs the retransmission control on a per-subblock basis. The
subblock retransmission method works as follows. When the receiver device
receives a PHY frame, it performs an error detection on a per-subblock basis and
requests the transmitter device to retransmit only the subblock(s) that contains an
error. Upon receiving the retransmission request, the transmitter device configures a
new PHY frame by appending a preamble to the specified subblock(s), and transmits
the new PHY frame. This method only requires retransmission of the subblock(s)
containing an error, instead of the entire PHY frame. Consequently, the throughput
reduction caused by the retransmission can be alleviated to a great extent. This
method also enables retransmission of data over a short period of time due to the
small size of the data, thus preventing reduction in the throughput of the entire
system.
[0009]
Patent Literature 1 describes a decoding method utilizing the maximum a
posteriori probability decoding. According to this decoding method, when decoding
transmission words that include known bits holding known values, a decoded word
candidate in which a known value in a part of a transmission word has been changed
to another value is excluded from the decoded word candidates. In Patent Literature
1, a synchronization byte of an MPEG-2TS packet is used as a known bit by way of
example.
[Citation List]
[Patent Literature]
[0010]
[Patent Literature 1]
JP Patent No. 4208017
[Non-Patent Literature]
[0011]
[Non-Patent Literature 1]
Hu et al. "On the Computation of the Minimum Distance of Low-Density
Parity Check codes." IEEE ICC2004.
[Non-Patent Literature 2]
IEEE 802.1 In, Draft 2.0
[Summary of Invention]
[Technical Problem]
[0012]
In order to apply the subblock retransmission method to a wireless
communication system conforming to the IEEE 802.1 In standard, it is possible to
use code blocks of an LDPC (Low-Density Parity-Check) code, which is used as an
error correction code, as subblocks. However, performing the retransmission control
while using LDPC code blocks as subblocks must require detecting whether each
LDPC code block contains an error.
[0013]
Well-known as an error detection method, a CRC code is an error detection
code that utilizes a cyclic code. The CRC code is widely used because of its high
error detection accuracy and its ability to easily achieve coding/decoding processing
with use of a shift register. However, in order to detect whether each LDPC code
block contains an error with use of the CRC code, redundant bits must be appended
to each LDPC code block for error detection purposes. This causes the problem of
throughput reduction.
[0014]
As one example of such a throughput reduction, the following describes the
case where a 32-bit CRC code is used for an LDPC code block having a code length
of 648 bits and a code rate of 1/2 as standardized in IEEE 802.1 In. The LDPC code
block having a code length of 648 bits and a code rate of 1/2 is made up of 324
information bits and 324 redundant bits. In order to apply a 32-bit CRC code to this
LDPC code block, 32 bits out of 324 information bits must be used for redundant
bits of the CRC code. In this case, a maximum of 292 information bits (324 - 32 =
292) can only be transmitted through one LDPC code block, and hence the
throughput is reduced by 10% (32/324 ~ 10%) compared to the case where all 324
bits are used as information bits.
[0015]
Meanwhile, there is also a method for performing an error detection on the
LDPC code block without appending error-detecting redundant bits. More
specifically, this method utilizes a parity check of an LDPC code. An LDPC code is
defined by a parity check matrix H. A code word c of the LDPC code always
satisfies the parity check equation shown in the following Expression 1.
[0016]
[Expression 1]
Hc = 0
[0017]
An error detection utilizing a parity check takes advantage of the above
relation to check whether an LDPC code block that has been subjected to processing
of error-correction decoding in the receiver device satisfies the parity check equation
He = 0. If the LDPC code block satisfies the parity check equation, then the LDPC
code block is judged as containing no error. If the LDPC code block does not satisfy
the parity check equation, then the LDPC code block is judged as containing an
error. As this method utilizes a parity check included the LDPC code, it is
advantageous in that error-detecting redundant bits need not be appended and a
throughput reduction is not caused.
[0018]
However, the problem with the error detection method utilizing the parity
check of the LDPC code is that, although the method can prevent a throughput
reduction, the method also gives rise to a false detection whereby an LDPC code
block that contains an error is judged as error-free at a rate of approximately 10"6 to
10'8. A description is now given of an error detection method utilizing a parity check
of an LDPC code with reference to FIG. 26.
[0019]
An LDPC decode unit 2610 performs LDPC decode processing on a
received code word cr and outputs the resultant decoded code word cd. A parity
check unit 2620 checks whether the decoded code word cd is a code word of the
LDPC code (this process is called a parity check). The parity check is performed by
checking whether the decoded code word cd satisfies the following Expression 2
with use of the parity check matrix H of the LDPC code.
[0020]
[Expression 2]
Hcd = 0
[0021]
The LDPC decode unit 2610 generally performs LDPC decode processing
using MAP decoding, such as sum-product decoding, or quasi-MAP decoding. The
MAP decoding and quasi-MAP decoding are performed based on an algorithm that
brings about a maximum a posteriori probability for each bit constituting a code
word. Therefore, a code word obtained as a result of the decoding is not necessarily
a code word of the LDPC code. When the decoded code word cd is not a code word
of the LDPC code, that is, when the decoded code word cd does not satisfy the above
Expression 2, inaccuracy of this decoding result can be detected.
[0022]
On the other hand, when the decoded code word cd is a code word of the
LDPC code, that is, when the decoded code word cd satisfies the above Expression 2,
there is a possibility that the decoded code word cd is an error-free code. Note that
the decoded code word cd being an error-free code means that the decoded code
word cd, which is obtained through the decoding performed by the receiver device, is
the same as the code word transmitted by the transmitter device. Also note that the
decoded code word cd being a code word of the LDPC code means that the decoded
code word cd satisfies the parity check of the LDPC code, i.e., the decoded code
word cd is judged as an error-free code word in the parity check.
[0023]
However, the decoded code word cd being a code word of the LDPC code is
not equal to the decoded code word cd being an error-free code word. This is
because, provided that the number of information bits in an LDPC code block is k,
there are 2k code words of the LDPC code, and the parity check only allows
checking whether the decoded code word cd matches any of the 2k code words.
Therefore, when an error detection utilizing the parity check is performed only by
itself, a code word that is both (i) a code word of the LDPC code and (ii) an
erroneous code word may be judged as containing no error. This is the reason for the
above-mentioned low detection accuracy.
[0024]
The following describes a rate at which a false detection takes place. Table
1 shows a block error rate and a false detection rate obtained by a computer
simulation for an LDPC code having a code length of 648 bits and a code rate of 1/2
as standardized in IEEE 802.1 In, in correspondence with Eb/NO (dB) in an AWGN
(Additive White Gaussian Noise) communication channel. The block error rate
denotes a rate at which a code word that has been subjected to LDPC decoding
(LDPC code block) contains an error. The false detection rate denotes a rate of
occurrence of an event where the code word that has been subjected to LDPC
decoding satisfies the parity check but is not an error-free code word. It is apparent
from Table 1 that, with respect to Eb/NO corresponding to a block error rate of 10"4
or less, a false detection takes place at a rate of approximately 10"6 to 10"7.
[0025]
[Table 1]
[0026]
The present invention has been made to solve the above problem associated
with an error detection method utilizing the parity check. The present invention aims
to provide (i) a coding device and a coding method capable of generating a code
word that can reduce a rate of false detection whereby an erroneous code word is
judged as an error-free code word with use of a parity check of an LDPC code, and
(ii) a decoding device and a decoding method capable of receiving and decoding the
code word generated by these coding device and coding method.
[Solution to Problem]
[0027]
In order to solve the above problem, a coding device pertaining to the
present invention is for coding an information bit sequence and outputting the coded
information bit sequence, and comprises: a determining unit operable to, in
accordance with an applied coding method that utilizes a quasi-cyclic code,
determine at least one bit position in the information bit sequence as an insertion
position into which a predetermined bit value is to be inserted, wherein a bit position
in a first code word which corresponds to the insertion position holds a different bit
value from a bit position in at least one second code word which corresponds to the
insertion position, the first code word being obtained by coding the information bit
sequence in accordance with the coding method, the second code word being judged
as an error-free code word by a parity check of the coding method but being
different from the first code word by the minimum Hamming distance; an inserting
unit operable to generate a pre-coding bit sequence by inserting the predetermined
bit value into the insertion position of the information bit sequence; and a coding
unit operable to generate a third code word by coding the pre-coding bit sequence in
accordance with the coding method.
[0028]
Also, a coding method pertaining to the present invention is for coding an
information bit sequence, comprising the steps of: (A) determining, in accordance
with an applied coding method that utilizes a quasi-cyclic code, at least one bit
position in the information bit sequence as an insertion position into which a
predetermined bit value is to be inserted, wherein a bit position in a first code word
which corresponds to the insertion position holds a different bit value from a bit
position in at least one second code word which corresponds to the insertion position,
the first code word being obtained by coding the information bit sequence in
accordance with the coding method, the second code word being judged as an
error-free code word by a parity check of the coding method but being different
from the first code word by the minimum Hamming distance; (B) generating a
pre-coding bit sequence by inserting the predetermined bit value into the insertion
position of the information bit sequence; and (C) generating a third code word by
coding the pre-coding bit sequence in accordance with the coding method.
[0029]
Also, a decoding device pertaining to the present invention comprises: a
decoding unit operable to generate a decoded bit sequence by performing decode
processing on a pre-decoding bit sequence input thereto, the decode processing
complying with an applied coding method that utilizes a quasi-cyclic code; a
checking unit operable to check whether the decoded bit sequence is judged as an
error-free code word by a parity check of the coding method; a judging unit operable
to judge whether at least one insertion position in the decoded bit sequence holds at
least one predetermined bit value, wherein according to the coding method, a bit
position in a first code word which corresponds to the insertion position is
predetermined to hold a different bit value from a bit position in a second code word
which corresponds to the insertion position, the first code word being generated by a
transmitter device according to the coding method, and the second code word being
judged as an error-free code word by the parity check but being different from the
first code word by the minimum Hamming distance; and an error detecting unit
operable to, in a case where the judging unit judges in the negative, judge that the
decoded bit sequent is erroneous even when a result of the checking by the checking
unit shows that the decoded bit sequence has been judged as the error-free code
word by the parity check.
[0030]
Also, a decoding method pertaining to the present invention comprises the
steps of: (A) generating a code word by decoding a received signal in accordance
with an applied coding method that utilizes a quasi-cyclic code; (B) checking
whether the code word generated in the step (A) is an error-free code word by a
parity check of the coding method; (C) judging whether at least one insertion
position in the generated code word holds at least one predetermined bit value,
wherein according to the coding method, a bit position in a first code word which
corresponds to the insertion position is predetermined to hold a different bit value
from a bit position in a second code word which corresponds to the insertion
position, the first code word being generated by a transmitter device according to the
coding method, and the second code word being judged as an error-free code word
by the parity check but being different from the first code word by the minimum
Hamming distance; and (D) in a case where the judgment in the step (C) is in the
negative, judging that the generated code word is erroneous even when a result of
the checking by the step (B) shows that the generated code word has been judged as
the error-free code word by the parity check.
[0031]
Also, a communication system pertaining to the present invention includes a
transmitter device and a receiver device. The transmitter device, which codes an
information bit sequence and transmits the coded information bit sequence,
comprises: a determining unit operable to, in accordance with an applied coding
method that utilizes a quasi-cyclic code, determine at least one bit position in the
information bit sequence as an insertion position into which a predetermined bit
value is to be inserted, wherein a bit position in a first code word which corresponds
to the insertion position holds a different bit value from a bit position in at least one
second code word which corresponds to the insertion position, the first code word
being obtained by coding the information bit sequence in accordance with the
coding method, the second code word being judged as an error-free code word by a
parity check of the coding method but being different from the first code word by
the minimum Hamming distance; an inserting unit operable to generate a pre-coding
bit sequence by inserting the predetermined bit value into the insertion position of
the information bit sequence; a coding unit operable to generate a third code word
by coding the pre-coding bit sequence in accordance with the coding method; and a
transmitting unit operable to transmit a wireless signal containing the third code
word. The receiver device comprises: a receiving unit operable to receive the
wireless signal; a decoding unit operable to generate a code word by decoding the
received signal in accordance with the coding method; a checking unit operable to
check whether the code word generated by the decoding unit is judged as an
error-free code word by the parity check of the coding method; a judging unit
operable to judge whether at least one insertion position in the generated code word
holds at least one predetermined bit value, wherein according to the coding method,
a bit position in a first code word which corresponds to the insertion position is
predetermined to hold a different bit value from a bit position in a second code word
which corresponds to the insertion position, the first code word being generated by
the transmitter device according to the coding method, and the second code word
being judged as an error-free code word by the parity check but being different from
the first code word by the minimum Hamming distance; and an error detecting unit
operable to, in a case where the judging unit judges in the negative, judge that the
generated code word is erroneous even when a result of the checking by the
checking unit shows that the generated code word has been judged as the error-free
code word by the parity check.
[Advantageous Effects of Invention]
[0032]
The above structure enables the coding device to, in accordance with an
applied coding method that utilizes quasi-cyclic codes, determine insertion positions
of an information bit sequence into which known bits are to be inserted. After the
known bits are inserted into the insertion positions, the encoding can be performed.
These insertion positions hold different bit values between (i) a code word obtained
by coding the information bit sequence to be transmitted as-is, and (ii) a code word
that is different from the code word (i) by the minimum Hamming distance but is
judged as an error-free code word by a parity check of the applied coding method.
Hence, by inserting bit values that are known to the transmitter device and the
receiver device into these insertion positions, the receiver device can detect an error
in a signal it has received if a decoded code word, which is obtained by decoding the
received signal, is different from the original code word transmitted by the
transmitter device by the minimum Hamming distance, even when the decoded code
word has changed from the original code word but has been judged as an error-free
code word by the parity check of the applied coding method. This can reduce the
rate at which a code word that is different from the original code word is falsely
judged as an error-free code word.
[Brief Description of Drawings]
[0033]
FIG. 1 shows one example of a parity check matrix pertaining to
Embodiment 1.
FIG. 2 shows code words that are different from an all-zero code word by
the minimum Hamming distance.
FIG. 3 is a conceptual diagram showing insertion positions of a code word
into which false detection check bits are to be inserted.
FIG. 4 shows a system structure of a communication system pertaining to
Embodiment 1.
FIG. 5 is a functional block diagram showing a functional structure of a
coding device included in a communication device pertaining to Embodiment 1.
FIG. 6 is a functional block diagram showing a functional structure of a
decoding device included in a communication device pertaining to Embodiment 1.
FIG. 7 is a functional block diagram showing a functional structure of an
error detection unit.
FIG. 8 shows an insertion position determination table associated with each
coding mode.
FIG. 9 is a flowchart showing coding operations performed by a coding
device pertaining to Embodiment 1.
FIGs. 10A to IOC are conceptual diagrams showing a concept of inserting
predetermined bit values into an information bit sequence.
FIG. 11 is a flowchart showing decoding operations performed by a
decoding device pertaining to Embodiment 1.
FIGs. 12A-1 to 12B-4 are conceptual diagrams relating to insertion of
shortening bits pertaining to Embodiment 2.
FIGs. 13A to 13E are conceptual diagrams relating to insertion of false
detection check bits pertaining to Embodiment 2.
FIG. 14 is a functional block diagram showing a functional structure of a
coding device pertaining to Embodiment 2.
FIG. 15 is a functional block diagram showing a functional structure of a
decoding device pertaining to Embodiment 2.
FIG. 16 is a flowchart showing coding operations performed by a coding
device pertaining to Embodiment 2.
FIG. 17 is a flowchart showing decoding operations performed by a
decoding device pertaining to Embodiment 2.
FIG. 18A shows the structure of an LDPC code word including block ID
bits. FIG. 18B is a conceptual diagram showing a concept of coding pertaining to
Embodiment 3. FIG. 18C shows the structure of a pre-coding bit sequence. FIG.
18D shows the structure of a code word pertaining to Embodiment 3.
FIG. 19 is a functional block diagram showing a functional structure of a
coding device pertaining to Embodiment 3.
FIG. 20 is a functional block diagram showing a functional structure of a
decoding device pertaining to Embodiment 3.
FIG. 21 is a functional block diagram showing a functional structure of an
error detection unit pertaining to Embodiment 3.
FIG. 22 is a flowchart showing coding operations performed by a coding
device pertaining to Embodiment 3.
FIG. 23 is a flowchart showing decoding operations performed by a
decoding device pertaining to Embodiment 3.
FIG. 24 is a functional block diagram showing an example of an alternative
structure for an error detection unit.
FIG. 25 is a data structure diagram showing a frame structure of signals
exchanged between communication devices.
FIG. 26 shows a concept of conventional LDPC decoding.
[Description of Embodiments]
[0034]
One aspect of the present invention is a first coding device for coding an
information bit sequence and outputting the coded information bit sequence,
comprising: a determining unit operable to, in accordance with an applied coding
method that utilizes a quasi-cyclic code, determine at least one bit position in the
information bit sequence as an insertion position into which a predetermined bit
value is to be inserted, wherein a bit position in a first code word which corresponds
to the insertion position holds a different bit value from a bit position in at least one
second code word which corresponds to the insertion position, the first code word
being obtained by coding the information bit sequence in accordance with the
coding method, the second code word being judged as an error-free code word by a
parity check of the coding method but being different from the first code word by
the minimum Hamming distance; an inserting unit operable to generate a pre-coding
bit sequence by inserting the predetermined bit value into the insertion position of
the information bit sequence; and a coding unit operable to generate a third code
word by coding the pre-coding bit sequence in accordance with the coding method.
[0035]
Another aspect of the present invention is a first decoding device
comprising: a decoding unit operable to generate a decoded bit sequence by
performing decode processing on a pre-decoding bit sequence input thereto, the
decode processing complying with an applied coding method that utilizes a
quasi-cyclic code; a checking unit operable to check whether the decoded bit
sequence is judged as an error-free code word by a parity check of the coding
method; a judging unit operable to judge whether at least one insertion position in
the decoded bit sequence holds at least one predetermined bit value, wherein
according to the coding method, a bit position in a first code word which
corresponds to the insertion position is predetermined to hold a different bit value
from a bit position in a second code word which corresponds to the insertion
position, the first code word being generated by a transmitter device according to the
coding method, and the second code word being judged as an error-free code word
by the parity check but being different from the first code word by the minimum
Hamming distance; and an error detecting unit operable to, in a case where the
judging unit judges in the negative, judge that the decoded bit sequent is erroneous
even when a result of the checking by the checking unit shows that the decoded bit
sequence has been judged as the error-free code word by the parity check.
[0036]
Yet another aspect of the present invention is a coding method for coding an
information bit sequence, comprising the steps of: (A) determining, in accordance
with an applied coding method that utilizes a quasi-cyclic code, at least one bit
position in the information bit sequence as an insertion position into which a
predetermined bit value is to be inserted, wherein a bit position in a first code word
which corresponds to the insertion position holds a different bit value from a bit
position in at least one second code word which corresponds to the insertion position,
the first code word being obtained by coding the information bit sequence in
accordance with the coding method, the second code word being judged as an
error-free code word by a parity check of the coding method but being different
from the first code word by the minimum Hamming distance; (B) generating a
pre-coding bit sequence by inserting the predetermined bit value into the insertion
position of the information bit sequence; and (C) generating a third code word by
coding the pre-coding bit sequence in accordance with the coding method.
[0037]
Yet another aspect of the present invention is a decoding method
comprising the steps of: (A) generating a code word by decoding a received signal in
accordance with an applied coding method that utilizes a quasi-cyclic code; (B)
checking whether the code word generated in the step (A) is an error-free code word
by a parity check of the coding method; (C) judging whether at least one insertion
position in the generated code word holds at least one predetermined bit value,
wherein according to the coding method, a bit position in a first code word which
corresponds to the insertion position is predetermined to hold a different bit value
from a bit position in a second code word which corresponds to the insertion
position, the first code word being generated by a transmitter device according to the
coding method, and the second code word being judged as an error-free code word
by the parity check but being different from the first code word by the minimum
Hamming distance; and (D) in a case where the judgment in the step (C) is in the
negative, judging that the generated code word is erroneous even when a result of
the checking by the step (B) shows that the generated code word has been judged as
the error-free code word by the parity check.
[0038]
Yet another aspect of the present invention is a communication system
including a transmitter device and a receiver device. The transmitter device, which
codes an information bit sequence and transmits the coded information bit sequence,
comprises: a determining unit operable to, in accordance with an applied coding
method that utilizes a quasi-cyclic code, determine at least one bit position in the
information bit sequence as an insertion position into which a predetermined bit
value is to be inserted, wherein a bit position in a first code word which corresponds
to the insertion position holds a different bit value from a bit position in at least one
second code word which corresponds to the insertion position, the first code word
being obtained by coding the information bit sequence in accordance with the
coding method, the second code word being judged as an error-free code word by a
parity check of the coding method but being different from the first code word by
the minimum Hamming distance; an inserting unit operable to generate a pre-coding
bit sequence by inserting the predetermined bit value into the insertion position of
the information bit sequence; a coding unit operable to generate a third code word
by coding the pre-coding bit sequence in accordance with the coding method; and a
transmitting unit operable to transmit a wireless signal containing the third code
word. The receiver device comprises: a receiving unit operable to receive the
wireless signal; a decoding unit operable to generate a code word by decoding the
received signal in accordance with the coding method; a checking unit operable to
check whether the code word generated by the decoding unit is judged as an
error-free code word by the parity check of the coding method; a judging unit
operable to judge whether at least one insertion position in the generated code word
holds at least one predetermined bit value, wherein according to the coding method,
a bit position in a first code word which corresponds to the insertion position is
predetermined to hold a different bit value from a bit position in a second code word
which corresponds to the insertion position, the first code word being generated by
the transmitter device according to the coding method, and the second code word
being judged as an error-free code word by the parity check but being different from
the first code word by the minimum Hamming distance; and an error detecting unit
operable to, in a case where the judging unit judges in the negative, judge that the
generated code word is erroneous even when a result of the checking by the
checking unit shows that the generated code word has been judged as the error-free
code word by the parity check.
[0039]
The above first coding device performs coding after inserting predetermined
bit values into positions that hold different values between (i) a code word generated
in accordance with an applied coding method that utilizes quasi-cyclic codes and (ii)
a code word that is judged as an error-free code word by a parity check of the coding
method but is different from the generated code word by the minimum Hamming
distance. A code word coded in the above manner enables the receiver device to
detect an erroneous code word that is judged as an error-free code word by the parity
check by checking the inserted bits different from predetermined bits.
[0040]
Yet another aspect of the present invention is a second coding device, which
is the first coding device further comprising a known bit appending unit operable to
append at least one known bit, which has been predetermined as a shortening bit of a
shortening code, to the information bit sequence when an information length of the
information bit sequence is shorter than an information length of an information bit
portion of a code word determined based on a codeword length and a code rate
pertaining to the coding method, wherein the inserting unit rearranges the
information bit sequence to which the known bit has been appended, so that the
known bit is inserted into the insertion position as the predetermined bit value.
[0041]
Yet another aspect of the present invention is a second decoding device,
which is the first decoding device further comprising an inserting unit operable to,
when a pre-decoding shortened code word from which one or more of the at least
one predetermined bit value have been removed by the transmitter device is input
thereto, generate the pre-decoding bit sequence by inserting at least one known bit
holding a known value into the input pre-decoding shortened code word.
[0042]
The above second coding device inserts known bits that have been
originally determined to be appended in order to adjust the information length of an
LDPC code block. This can suppress reduction in the throughput of data transfer
caused by the first coding device inserting predetermined bits.
[0043]
Yet another aspect of the present invention is a third coding device, which is
the first coding device further comprising a removing unit operable to remove one or
more of the at least one known bit appended by the known bit appending unit from
the third code word.
[0044]
The above third coding device can reduce the size of information to be
transmitted by removing known bits. This way, the throughput of data transfer can
be increased.
[0045]
Yet another aspect of the present invention is a fourth coding device, which
is the first coding device wherein the third code word generated by the coding unit
includes a block ID for identifying the third code word among other third code
words, and the inserting unit inserts a bit value of a bit representing the block ID
into the insertion position as the predetermined bit value.
[0046]
Yet another aspect of the present invention is a third decoding device, which
is the first decoding device wherein the judging unit judges whether the insertion
position in the decoded bit sequence holds a bit value that can be predicted from a
predetermined bit value of another decoded bit sequence, instead of judging whether
the insertion position in the decoded bit sequence holds the at least one
predetermined bit value.
[0047]
The above fourth coding device uses, as known bits to be inserted, block
IDs that have been originally determined to be appended and that determine the
order of code words. This can suppress reduction in the throughput of data transfer
caused by the first coding device inserting predetermined bits.
[0048]
The above third decoding device can detect whether a target block ID is
erroneous based on a block ID obtained from a previous or next code block. This
improves the rate of false detection whereby an erroneous code word is judged as
error-free.
[Embodiment 1]
The following describes one embodiment of a coding device and a decoding
device pertaining to the present invention with reference to the drawings.

Prior to explaining a coding device and a decoding device pertaining to the
present invention, concepts of the present invention are first described below.
[0049]
As has been exhibited in the above Technical Problem section, a coding
device and a decoding device pertaining to the present embodiment increase the rate
of prevention of false detection whereby an erroneous code word that satisfies a
parity check is judged as an error-free code word as a result of an error detection.
[0050]
As mentioned above, in a case where a sufficient SNR (Signal-to-Noise
Ratio) is preserved in the communication channel, the rate at which the decoded
code word cd becomes both a code word of the LDPC code and an erroneous code
word is approximately 10"6 to 10"8. Note that this rate varies depending on the parity
check matrix H defining the LDPC code.
[0051]
Below is a discussion on a decoded code word that contains an error but is
mistakenly judged by the decoding device as an error-free code word. More
specifically, the following discussion focuses on how likely it is for the decoded
code word cd, which is a code word of the LDPC code but is an erroneous code word,
to be judged as one of (2k - 1) code words that are code words of the LDPC code but
are erroneous code words. The decoded code word cd is obtained by performing
LDPC decoding on a received code word cr. The received code word cr is obtained
by the receiver device receiving a signal generated based on a corresponding
error-free code word. Provided that a sufficient SNR is preserved during
communication, even if the corresponding error-free code word changes during
communication, the amount of such a change in the corresponding error-free code
word is considered minimum. More specifically, when the decoded code word cd is a
code word of the LDPC code but is an erroneous code word, there is a high
possibility that the decoded code word cd is close to the corresponding error-free
code word in terms of the Hamming distance. Especially, there is an extremely high
possibility that the decoded code word cd is a code word that is different from the
corresponding error-free code word by the minimum Hamming distance (hereinafter
referred to as a "minimum-distance code word"). Therefore, by checking whether
the decoded code word cd is the minimum-distance code word in relation to the
corresponding error-free code word in addition to performing the conventional
parity check, the error detection accuracy can be improved to a practically sufficient
level. Note that the operation of "checking whether or the decoded code word cd is
the minimum-distance code word" is herein referred to as a minimum-distance code
word check. In the present Description, the minimum-distance code word denotes a
code word that is different from the corresponding error-free code word by a
Hamming distance of one or more.
[0052]
The following describes the minimum-distance code word check by taking
one specific example.
[0053]
In the following description of the minimum-distance code word check, an
LDPC code having a code length of 648 bits and a code rate of 1/2 as standardized
in IEEE 802.1 In is used as an example.
[0054]
FIG. 1 shows a parity check matrix Hb648 of an LDPC code having a code
length of 648 bits and a code rate of 1/2 as standardized in IEEE 802.1 In. The parity
check matrix Hb648 shown in FIG. 1 is a 12 * 24 matrix.
[0055]
Of all the entries in Hb648, a dash "-" indicates a 27 x 27 zero matrix.
[0056]
Also, of all the entries in Hb648, each entry expressed by an integer p
(where 0 < p) indicates a matrix obtained by performing a rightward cyclic shift of a
27 x 27 identity matrix p times. For example, an entry expressed by p having a value
0 is a 27 x 27 identity matrix, and an entry expressed by p having a value 22 is a
matrix obtained by performing a rightward cyclic shift of the 27 x 27 identity matrix
22 times.
[0057]
The parity check matrix Hb648 is made up of the above-described entries.
When expanded, the parity check matrix Hb648 is actually a 324 x 648 matrix
(defined herein as Hf648).
[0058]
Columns of Hf648 correspond to 648 bits constituting a code word,
respectively. The 1st to 324th columns make an information bit portion corresponding
to information bits, whereas the 325th to 648th columns make a parity bit portion. In
addition, each row of Hf648 corresponds to a parity check equation. In each row, a
summation of bits belonging to any column showing an entry with a value 1 (i.e., a
result of performing an exclusive or operation on such bits) is 0. An LDPC code
defined by Hb648 is referred to as a quasi-cyclic LDPC (Quasi-Cyclic Low-Density
Parity-Check) code.
[0059]
By using the minimum distance retrieval method disclosed in Non-Patent
Literature 1, the following can be retrieved: (i) the minimum Hamming distance
between an original error-free code word and a corresponding erroneous code word
of the LDPC code defined by Hb648 is 15; and (ii) there are 27 minimum-distance
code words. The minimum distance being 15 means that, between an original
error-free code word and a corresponding minimum-distance code word, 15 bits out
of 648 bits show different values, whereas the remaining 633 bits show the same
values. For simplicity, the following description is provided under the assumption
that an error-free code word is a 648-bit all-zero code word. The 648-bit all-zero
code word satisfies He = 0 and therefore is a code word of Hb648. It should be
mentioned that as an LDPC code is a linear code, generality of the following
description would not be lost even when the error-free code word is other than the
all-zero code word.
[0060]
FIG. 2 shows a list of information bit portions of the 27 minimum-distance
code words. Between the all-zero code word and each minimum-distance code word,
15 bits are different—i.e., there is a Hamming distance of 15 therebetween.
However, as for the information bit portion of each minimum-distance code word, 4
bits out of the entire 324 bits are different. It can be seen from FIG. 2 that the 27
minimum-distance code words have relevance to one another. By "relevance" it
means that the 4 bits different from their counterparts in the corresponding error-free
code word (in this example, the all-zero code word) are distributed as follows: of the
4 bits, 2 bits are in the 4th column, 1 bit is in the 10th column, and 1 bit is in the 12th
column of Hb648 as shown in FIG. 2.
[0061]
Furthermore, when the identity matrix is expanded, each of the 27
minimum-distance code words has the following bit arrangement: provided n = 1 to
27, values of the nth and (n + l)%27th bits in the 4th column, the (n + 17)%27th bit in
the 10th column, and the (n + l)%27th bit in the 12th column of Hb648 are different
from their counterparts in the corresponding error-free code word. The above
relevance arises from the following reason. Since an LDPC code conforming to
IEEE 802.1 In shown in FIG. 2 is a quasi-cyclic code (QC-LDPC code), a bit
sequence obtained by cycling a given code word in units of the identity matrix size
(i.e., in units of 27 bits) is also a code word, and the Hamming weight thereof
remains unchanged. Note that the symbol "%" used herein denotes a modulo
operation.
[0062]
The following describes a method of detecting whether the decoded code
word cd satisfying the parity check equation is a minimum-distance code word with
n = 1 (i.e., a code word that is different from a corresponding error-free code word
in values of the 82nd, 83rd, 261st and 299th bits in the information bit portion thereof)
by taking advantage of the above-described relevance between the error-free code
word and each of the 27 minimum-distance code words.
[0063]
The transmitter device generates an error-free code word including at least
one known bit (i.e., a bit holding a known value), the stated at least one bit
corresponding to one of four bits that are included in a corresponding
minimum-distance code word with n = 1 and hold different values from their
counterparts in the error-free code word. It should be noted that information
indicating which bit is set as a known bit and information indicating what value is
held by the known bit are shared between the transmitter device and the receiver
device ahead of time. With such a known bit inserted in the information bit sequence,
the receiver device can judge that the decoded code word cd is not a
minimum-distance code word with n = 1 when the value of the known bit in the
decoded code word cd is the same as the known value preset by the transmitter
device. On the other hand, when the value of the known bit in the decoded code
word cd is different from the known value preset by the transmitter device, the
receiver device can judge that there is a possibility that the decoded code word cd is
at least not an error-free code word but is a minimum-distance code word with n = 1.
[0064]
The receiver device judges the decoded code word cd as containing no error
only in a case where the decoded code word cd satisfies the parity check and its
known bit holds a known value. In other cases, the receiver device judges the
decoded code word cd as containing an error. This makes it possible to eliminate the
possibility of false detection whereby the decoded code word cd that is a
minimum-distance code word with n = 1 is judged as containing no error.
[0065]
The same goes for other minimum-distance code words with n = 2 to 27.
That is, out of four bits of each minimum-distance code word that hold different
values from their counterparts in the corresponding error-free code word, at least one
bit is set as a known bit. This makes it possible to eliminate the possibility of false
detection whereby the decoded code word cd that is one of the minimum-distance
code words with n = 1 to 27 is judged as containing no error. To be more specific,
the receiver device should be configured as follows. The receiver device judges the
decoded code word cd as containing no error only in a case the following conditions
are both satisfied: (i) the decoded code word cd satisfies the parity check; and (ii) all
known bits in the decoded code word cd hold known values preset by the transmitter
device. The receiver device judges the decoded code word cd as containing an error
in any other case.
[0066]
As set forth above, in an information bit portion of every minimum-distance
code word, at least one bit out of all bits holding different values from their
counterparts in a corresponding error-free code word is preset as a known bit. This
way, when the decoded code word cd is not an error-free code word but is a
minimum-distance code word, the receiver device is prevented from making a false
detection whereby the decoded code word cd is judged as containing no error.
Hereinafter, a known bit that holds a known value for the purpose of avoiding the
above-described false detection is referred to as a false detection check bit.
[0067]
In the case of an LDPC code conforming to the IEEE 802.1 In standard with
a codeword length of 648 bits and a code rate of 1/2, the 4th column in the parity
check matrix Hb648 includes bits holding different values between an error-free
code word and a corresponding minimum-distance code word. Hence, by setting one
bit that is in a position corresponding to the fourth column of the parity check matrix
Hb648 as a known bit, two minimum-distance code words can be detected at once.
[0068]
Therefore, as shown in FIG. 3, in the case of an LDPC code conforming to
IEEE 802.1 In with a codeword length of 648 bits and a code rate of 1/2, the 14 bits
corresponding to odd-numbered columns out of all the 27 columns shown in the 4th
column of the parity check matrix Hb648 are set as known bits x. This arrangement
allows checking whether the decoded code word Cd is a minimum-distance code
word, no matter which one of the 27 minimum-distance code words the decoded
code word cd represents. Note that a false detection check bit x may hold any value
as long as that value is known to the transmitter device and the receiver device, i.e.,
as long as that value is shared between the transmitter device and the receiver device
ahead of time.
[0069]
In the present example, all 27 code words having the minimum Hamming
distance can be excluded by setting the 14 bits corresponding to the odd-numbered
columns out of all the 27 columns shown in the 4th column of Hb648 as false
detection check bits. However, as long as at least one of these 14 bits is set as a false
detection check bit, it is possible to eliminate the possibility of falsely detect at least
two code words having small Hamming distances as error-free code words, and
therefore to reduce the rate of false detection. Hence, in order to increase the
throughput of information transmission, only part of these 14 bits may be set as a
known bit(s). Meanwhile, by setting at least one bit in columns other than the 4th
column of Hb648 (i.e., in the 10th and 12th columns) as a false detection check bit
(i.e., known bit), it is possible to eliminate the possibility of falsely detecting at least
one code word having the small Hamming distance as an error-free code word.

A description is now given of a transmitter device and a receiver device that
can prevent the above-described false detection.
[0070]
FIG. 4 shows an example of a system structure of a communication system
that uses a coding device and a decoding device. The communication system
includes communication devices 10a and 10b that each function as a transmitter
device or a receiver device. The communication devices 10a and 10b perform
communication via a network 20. Each of the communication devices 10a and 10b
includes a coding device 100 and a decoding device 200.
[0071]
As shown in FIG. 4, in the communication system, a plurality of
communication devices are connected with one another via the network. Each one of
the communication device includes a coding device and a decoding device.
Examples of a communication device include a cell phone and a base station device.
[0072]
A communication device transmits information to another communication
device after performing the coding processing shown in the present embodiment (the
specifics are described later). The stated another communication device receives a
transmitted signal and performs the decode processing shown in the present
embodiment (the specifics are described later).
[0073]
In the present Description, the communication processing that does not
involve the coding processing and the LDPC decoding performed by communication
devices (for example, AGC, ADC (Analog-to-Digital Convert), DAC
(Digital-to-Analog Convert), FFT (Fast Fourier Transform), IFFT (Inverse Fast
Fourier Transform), and quadrature detection) is performed according to
conventional methods. Hence, the details of such processing are omitted. It goes
without saying that each communication device, as the name suggests, includes
other circuits for signal transmission/reception. The details of such circuits are
omitted as well.
[0074]
FIG. 5 is a functional block diagram showing the functional structure of the
coding device included in each communication device.
[0075]
As shown in FIG. 5, the coding device 100 includes an LDPC coding
parameter storage unit 101, a false detection check bit control unit 102, a false
detection check bit insertion unit 103, and an LDPC coding unit 104.
[0076]
The LDPC coding parameter storage unit 101 has the function of storing
various types of parameters required for the LDPC coding in accordance with
coding mode with which the LDPC coding is to be performed, as well as the
function of notifying such coding mode with which the LDPC coding is to be
performed to the false detection check bit control unit 102 and the LDPC coding unit
104. Coding mode is determined based on the codeword length and code rate. In the
case of IEEE 802.1 In, three types of codeword lengths are defined, namely 648 bits,
1296 bits, and 1944 bits, and four types of code rates are defined, namely 1/2, 2/3,
3/4, and 5/6. That is to say, there is a total of 12 coding modes. Each communication
device selects one of the three types of codeword lengths and one of the four types
of code rates. The coding device therein performs the coding in accordance with the
selected codeword length and code rate. The various types of parameters include, for
example, a codeword length N, an information bit length K, a redundant bit length
M, a parity check matrix H, the number D of false detection check bit, false
detection check bit insertion positions p, etc. It should be noted that a transmitter
device includes information indicating one or more of the above parameters
(hereinafter referred to as coding mode information) into SIG and H-SIG, which are
transmitted as the preamble signal of the above-described PHY frame. This way, the
receiver device can analyze the SIG and H-SIG to obtain the coding mode.
[0077]
The LDPC coding parameter storage unit 101 identifies the coding mode
with which the coding is to be performed in accordance with the coding mode
information specified by the CPU or the like of the communication device, by using
an insertion position determination table that defines false detection check bit
insertion positions, which are one of the aforementioned various parameters (the
details of the table are described later). Then, the LDPC coding parameter storage
unit 101 notifies information relating to the specified coding mode to the false
detection check bit control unit 102 and the LDPC coding unit 104. The LDPC
coding parameter storage unit 101 notifies the false detection check bit control unit
102 of the number of false detection check bits to be inserted as well as insertion
positions of the false detection check bits. The LDPC coding parameter storage unit
101 also notifies the LDPC coding unit 104 of a parity check matrix H to be used.
[0078]
The false detection check bit control unit 102 generates the necessary
number of false detection check bits based on the information notified by the LDPC
coding parameter storage unit 101. Note that the false detection check bits may hold
any values as long as the values are known to both of the transmitter device and the
receiver device. For simplicity, the following description will be given assuming
that all the false detection check bits hold a value 0. The false detection check bit
control unit 102 notifies the false detection check bit insertion unit 103 of the false
detection check bit insertion positions p of the corresponding coding mode, together
with the generated false detection check bits.
[0079]
The false detection check bit insertion unit 103 generates a pre-coding bit
sequence by inserting, into the false detection check bit insertion positions specified
by the false detection check bit control unit 102, the false detection check bits that
have been notified by the false detection check bit control unit 102 as holding
predetermined bit values. Then, the false detection check bit insertion unit 103
transfers the generated pre-coding bit sequence to the LDPC coding unit 104.
[0080]
The LDPC coding unit 104 has the function of (i) generating a code word
by executing the LDPC coding processing on the information bit sequence into
which the false detection check bits have been inserted—i.e., the pre-coding bit
sequence transferred from the false detection check bit insertion unit
103—according to the parity check matrix H notified by the LDPC coding
parameter storage unit 101, and (ii) outputting the generated code word.
[0081]
This code word is modulated by a transmission circuit and then transmitted
via an antenna.
[0082]
FIG. 6 is a functional block diagram showing the functional structure of the
decoding device included in each communication device.
[0083]
As shown in FIG. 6, the decoding device 200 includes an LDPC coding
parameter storage unit 201, an LDPC decode unit 202, a false detection check bit
control unit 203, an error detection unit 204, and a false detection check bit removal
unit 205.
[0084]
The LDPC coding parameter storage unit 201 stores therein sets of
parameters required for LDPC decode processing and error detection processing,
such as a code length N, an information bit length K, a redundant bit length M, a
parity check matrix H, the number D of false detection check bits, and false
detection check bit insertion positions p. Note that the number of sets of parameters
stored in the LDPC coding parameter storage unit 201 is equal to the number of
coding modes that could be implemented by the decoding device 200. The LDPC
coding parameter storage unit 201 is notified of the coding mode to be implemented
via, for example, the coding mode information that specifies at least the codeword
length and code rate and that is obtained by decoding a signal received via a
reception circuit (not illustrated) of the communication device and analyzing the
SIG and H-SIG included in the preamble signal of the decoded signal.
[0085]
Also, the LDPC coding parameter storage unit 201 transfers the parity
check matrix H and false detection check bit insertion positions p corresponding to
the notified coding mode information to the false detection check bit control unit
203.
[0086]
The LDPC decode unit 202 performs the LDPC decode processing on the
received code word input to the decoding device 200. The LDPC decode unit 202
performs arbitrary decode processing for decoding the LDPC code, such as
sum-product decoding and min-sum decoding. The LDPC decode unit 202 transfers
the code word obtained as a result of the LDPC decode processing to the false
detection check bit removal unit 205 and the error detection unit 204. It should be
mentioned here that the present invention may use any method for LDPC decoding.
[0087]
The false detection check bit control unit 203 notifies the error detection
unit 204 of the false detection check bit insertion positions p notified from the LDPC
coding parameter storage unit 201, as well as bit values held by the inserted false
detection check bits.
[0088]
The false detection check bit control unit 203 also notifies the false
detection check bit removal unit 205 of the false detection check bit insertion
positions p notified from the LDPC coding parameter storage unit 201.
[0089]
The error detection unit 204 has the function of (i) performing an error
detection on the code word obtained as a result of the LDPC decoding, and (ii)
outputting the error detection result as a result of detection by the decoding device
200.
[0090]
FIG. 7 shows a detailed functional structure of the error detection unit 204.
As shown in FIG. 7, the error detection unit 204 includes a parity check subunit 211
and a false detection check subunit 212.
[0091]
The parity check subunit 211 checks whether a decoded code word cd that
has been input and the parity check matrix H of the LDPC code that has been
provided as control information satisfy the parity check Hcd = 0. The parity check
subunit 211 outputs a result of the parity check in format that can be interpreted by
software or hardware. For example, the parity check subunit 211 outputs "0" when
the parity check is satisfied, and outputs "1" when the parity check is not satisfied.
[0092]
The decoded code word cd, the result of the parity check, and the control
information indicating the number, values and insertion positions of false detection
check bits are input to the false detection check subunit 212.
[0093]
When the result of the parity check shows "0", i.e., when the parity check
subunit 211 has judged that the parity check is not satisfied, the false detection
check subunit 212 checks whether the false detection check bits that have been
inserted by the transmitter device into the information bit portions are known bits.
The false detection check subunit 212 outputs a result of the false detection check in
format that can be interpreted by software or hardware. For example, the false
detection check subunit 212 outputs "0" when no false detection has been confirmed,
and "1" when a false detection has been confirmed. The result of the false detection
check output from the false detection check subunit 212 is also a result of the error
detection output from the error detection unit 204. On the other hand, when the
result of the parity check shows "1", i.e., when the parity check subunit 211 has
judged that the parity check is not satisfied, it is obvious that the decoded code word
cd is not a code word of the LDPC code and is thus erroneous. In this case, there is
no need to perform the false detection check. Therefore, in this case, the false
detection check subunit 212 outputs "1" as a result of the false detection check
without performing the false detection check.
[0094]
Returning to FIG. 6, the false detection check bit removal unit 205 removes
D false detection check bits and M redundant bits based on the false detection check
bit insertion positions p that are included in the code word subjected to the LDPC
decoding and are notified from the false detection check bit control unit 203.
Thereafter, the false detection check bit removal unit 205 outputs only a J-bit
decoded information bit sequence as the output from the decoding device 200.

A description is now given of an insertion position determination table
corresponding to the coding mode information stored in the LDPC coding parameter
storage unit 101. The insertion position determination table is information showing
association between a mode number, a codeword length, a code rate, and insertion
positions. The insertion position determination table is prestored in both the
transmitter device and the receiver device.
[0095]
The mode number is set by the coding device for convenience in order to
manage each mode.
[0096]
The codeword length is information indicating a code length of a code word
to be generated. In the case of IEEE 802.1 In, the codeword length is one of 648 bits,
1296 bits, and 1944 bits.
[0097]
The code rate is information indicating the ratio of an information bit
sequence to the code word to be generated. In the case of IEEE 802.1 In, the code
rate is one of 1/2, 2/3, 3/4 and 5/6.
[0098]
The false detection check bit insertion positions are information indicating
insertion positions in which false detection check bits should be inserted for
preventing a false detection in the code word to be generated based on the
corresponding codeword length and code rate. The false detection check bit insertion
positions are also information that defines the orders of bits in which the false
detection check bits should be inserted counting from the beginning of the
information bit portion of the code word to be generated.
[0099]
As the transmitter device holds the insertion position determination table, it
can determine the insertion positions corresponding to the codeword length and code
rate indicated by the coding mode information, and insert the false detection check
bits for preventing a false detection.
[0100]
Furthermore, as the receiver device also holds the insertion position
determination table, it can identify the positions in which the false detection check
bits have been inserted, and detect whether the received code word is an error-free
code word by checking if the bit values at the identified positions are the same as
predetermined bit values prestored therein. If the bit values at the identified
positions are not the same as the predetermined bit values, then it follows that the
received code word is different from a corresponding code word generated by the
transmitter device. In this case, a failure in reception can be detected even if the
received code word is judged as an error-free code word as a result of the parity
check.
[0101]
It should be noted that the insertion position determination table described
herein may be included as part of information relating to the parameters stored in the
LDPC coding parameter storage units 101 and 201. Alternatively, the insertion
position determination table may be associated with information indicating the
number of false detection check bits to be inserted, an information length of an
information bit sequence, etc., for each coding mode.

Described below are coding operations performed by the coding device 100
pertaining to the present embodiment with reference to the flowchart of FIG. 9.
[0102]
First, the coding device 100 receives coding mode information as input
(step S901). This coding mode information is notified from the CPU (not illustrated)
or the like of the communication device and has been determined in accordance with,
for instance, the communication standard to adopt.
[0103]
The LDPC coding parameter storage unit 101 identifies false detection
check bit insertion positions p based on the codeword length and code rate specified
by the coding mode information, and notifies the false detection check bit control
unit 102 of the false detection check bit insertion positions p and the number D of
false detection check bits to be inserted. Additionally, the LDPC coding parameter
storage unit 101 notifies the LDPC coding unit 104 of the parity check matrix H
specified by the coding mode information (step S902).
[0104]
The false detection check bit control unit 102 generates as many
predetermined known bits as there are D false detection check bits to be inserted,
and notifies the false detection check bit insertion unit 103 of the generated known
bits and the false detection check bit insertion positions p (step S903).
[0105]
After the false detection check bit control unit 102 has notified the false
detection check bit insertion unit 103 of the false detection check bit insertion
positions p and the bit values to be inserted as false detection check bits, the false
detection check bit insertion unit 103 acquires, from a buffer or the like (not
illustrated) of the communication device storing therein information to be
transmitted, a partial information bit sequence of an information bit portion of a
code word to be transmitted, the partial information bit sequence having a shorter
information length than the stated information bit portion by the number of false
detection check bits to be inserted.
[0106]
Then, the false detection check bit insertion unit 103 sequentially inserts the
notified known bits into the positions specified by the false detection check bit
insertion positions p, starting from the beginning of the acquired partial information
bit sequence. The false detection check bit insertion unit 103 generates a pre-coding
bit sequence through the above procedure, and transfers the generated pre-coding bit
sequence to the LDPC coding unit (step S904).
[0107]
The LDPC coding unit 104 generates a code word by coding the pre-coding
bit sequence transferred from the false detection check bit insertion unit 103 in
accordance with the parity check matrix H notified from the LDPC coding
parameter storage unit 101, outputs the generated code word (step S905), and ends
the processing. Code words are generated by repeating the above operations as long
as data exists in the buffer storing therein the information to be transmitted.
[0108]
A transmission circuit provided in the communication device performs
subcarrier modulation, FFT, D/A conversion, quadrature modulation, etc. on the
code word generated by the LDPC coding unit 104. Thereafter, the code word is
transmitted via an antenna.
[0109]
Described below is one specific example of execution of coding with
reference to the flowchart of FIG. 9.
[0110]
More specifically, the following describes operations performed by the
coding device 100 with use of an LDPC code defined by the parity check matrix
Hb648 shown in FIG. 1 and having a code length of 648 bits and a code rate of 1/2
as specified in IEEE 802.1 In.
[0111]
The transmitter device inputs coding mode information and an information
bit sequence to be coded to the coding device 100. Here, the input information bit
sequence is a 310-bit binary sequence, which is shorter than the information block
length K (324 bits) by the number D of false detection check bits (14 bits). That is, J
= K-D = 310. The coding mode information is input to the LDPC coding
parameter storage unit 101. The LDPC coding parameter storage unit 101 sets
parameters required for the LDPC coding in the false detection check bit control unit
102 and the LDPC coding unit 104 in accordance with the input coding mode
information. Examples of the parameters to be set in the false detection check bit
control unit 102 include the number D of false detection check bits (14 bits), and
false detection check bit insertion positions p (the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th,
96*, 98th, 100th, 102nd, 104th, 106th, and 108th bits). Examples of the parameters to be
set in the LDPC coding unit 104 include the code length N (648 bits), the
information bit length K (324 bits), the redundant bit length M (324 bits), the parity
check matrix H (or an address of the memory storing therein information relating to
the parity check matrix H), and in a case where different processing is performed for
each coding mode, a signal indicating which processing is to be performed.
[0112]
The false detection check bit control unit 102 generates 14 false detection
check bits and transfers the same to the false detection check bit insertion unit 103.
The false detection check bit insertion unit 103 allocates the false detection check
bits respectively to the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th, 100th, 102nd,
104th, 106th, and 108th bits of the 324-bit sequence to be transferred to the LDPC
coding unit 104, as specified by the false detection check bit insertion positions p.
The false detection check bit insertion unit 103 also allocates 310 information bits
contained in the information bit sequence input from the coding device 100 to the
remaining 310 positions. Then, the false detection check bit insertion unit 103
transfers the information bit sequence into which the false detection check bits have
been inserted (a pre-coding information bit sequence) to the LDPC coding unit 104.
[0113]
The LDPC coding unit 104 performs LDPC coding processing on the input
pre-coding information bit sequence, generates a 324-bit redundant bit sequence,
combines the generated 324-bit redundant bit sequence with the 324-bit coded
information bit sequence, and finally outputs a result of the combination as a 648-bit
code word.
[0114]
With reference to FIGs. 10A to IOC, below is a schematic description of a
concept of coding pertaining to the present Embodiment 1 performed in accordance
with the coding method shown in FIG. 9.
[0115]
First of all, let us assume that an information bit sequence shown in FIG.
10A is to be transmitted.
[0116]
In accordance with the coding mode information, the false detection check
bit control unit 102 determines insertion positions in which false detection check bits
are to be inserted, and inserts the false detection check bits, which are known bits,
into the determined insertion positions. In the present example, the number D of
false detection check bits to be inserted is four (i.e., D = 4).
[0117]
In this case, as shown in FIG. 10B, each false detection check bit is inserted
between two information bit sequences. As a result, a pre-coding bit sequence is
generated.
[0118]
The LDPC coding unit 104 generates code word bits by appending a
redundant bit sequence based on the parity check matrix of the QC-LDPC code
(parity bit sequence) to the pre-coding bit sequence shown in FIG. 10B.
[0119]
Note that in FIGs. 10A to IOC, K bits, M bits and N bits represent the
information length of the information bit sequence(s) in the code word, the length of
the redundant bit sequence in the code word, and the length of the code word,
respectively. The codeword length, namely N bits, is determined based on the
coding mode information. K and M are determined based on the code rate included
in the coding mode information.
[0120]
The coding device pertaining to the present embodiment generates a code
word in the above manner. Inserting false detection check bits between information
bit sequences reduces the throughput of data transfer. However, the amount of
information inserted in this manner is smaller than the amount of information
inserted by appending a CRC code (for example, as shown in FIG. 2, 14 false
detection check bits are inserted to remove all code words having a close Hamming
distance). Therefore, the throughput of data transfer resulting from inserting false
detection check bits between information bit sequences is still higher than that
resulting from appending of the CRC code.
[0121]
The following describes operations of the decoding device 200 pertaining to
the present embodiment with reference to the flowchart of FIG. 11.
[0122]
The decoding device 200 receives coding mode information showing a
codeword length and a code rate by analyzing a preamble of the signal received by
the communication device (step SI 101).
[0123]
Upon receiving the coding mode information, the LDPC coding parameter
storage unit 201 notifies the false detection check bit control unit 203 of the false
detection check bit insertion positions p and the number of false detection check bits,
which are identified by the coding mode information. The LDPC coding parameter
storage unit 201 also notifies the LDPC decode unit 202 of the parity check matrix
H corresponding to the coding mode information (step SI 102).
[0124]
The false detection check bit control unit 203 notifies the error detection
unit 204 and the false detection check bit removal unit 205 of the notified false
detection check bit insertion positions p as well as values of known bits inserted as
the false detection check bits (step SI 103).
[0125]
The LDPC decode unit 202 performs LDPC decoding in accordance with
the parity check matrix H notified from the LDPC coding parameter storage unit 201,
and notifies the error detection unit 204 and the false detection check bit removal
unit 205 of the resultant decoded code word (step SI 104).
[0126]
The error detection unit 204 detects whether the bit values inserted in the
notified false detection check bit insertion positions are the same as values of the
notified known bits. If the result of the detection is affirmative, then the error
detection unit 204 judges the decoded code word as an error-free code word and
outputs "0" as an error detection result. On the other hand, if the result of the
detection is negative, then the error detection unit 204 judges the decoded code work
as an erroneous code word and outputs "1" as an error detection result (step SI 105).
[0127]
The false detection check bit removal unit 205 removes the bit values at the
notified false detection check bit insertion positions and the appended redundant bits
from the decoded code word, outputs the resultant decoded information bit sequence
(step SI 106), and ends the decode processing.
[0128]
In the above manner, the decoding device 200 can detect (i) whether the
received and decoded code word satisfies the parity check, and (ii) whether the
received and decoded code word is different from a corresponding code word
generated by the transmitter device with a close Hamming distance therebetween.
The above procedure allows lowering the rate of false detection whereby a code
word that is different from a corresponding code word generated by the transmitter
device with a close Hamming distance therebetween is judged as an error-free code
word.
[0129]
The following explains one specific example of the above-described error
detection and decoding.
[0130]
More specifically, the following is a description of operations performed by
the decoding device 200 with use of an LDPC code defined by the parity check
matrix Hb648 shown in FIG. 1 and having a code length of 648 bits and a code rate
of 1/2 as specified in EEEE 802.1 In.
[0131]
Before the received code word is input, the coding mode information is
input to the decoding device 200. The LDPC coding parameter storage unit 201
transfers LDPC coding parameters which are stored in correspondence with the
received coding mode information to the LDPC decode unit 202 and the false
detection check bit control unit 203. Examples of the parameters to be set in the
false detection check bit control unit 203 include the number of false detection check
bits (14) and false detection check bit insertion positions p (the 82nd, 84th, 86th, 88th,
90th, 92nd, 94th, 96th, 98th, 100th, 102nd, 104th, 106th, and 108th bits). Examples of the
parameters to be set in the LDPC decode unit 202 include the code length (N) (648
bits), the information bit length K (324 bits), the redundant bit length M (324 bits),
the parity check matrix H (or an address of the memory storing therein information
relating to the parity check matrix H), and in a case where different processing is
performed for each coding mode, a signal indicating which processing is to be
performed.
[0132]
After the parameters to be used are set, the LDPC decode processing begins
for the received code word. The received 648-bit code word is input to the LDPC
decode unit 202. At this time, in the received code word, one bit may carry one-bit
information (hard-decision decoding), or one bit may carry multiple-bit information
(soft-decision decoding). In the case of soft-decision decoding, log-likelihood ratios
of quantized code word bits may be used as the received code word. The LDPC
decode unit 202 performs LDPC decode processing on the received code word, and
transfers the resultant code word to the error detection unit 204 and the false
detection check bit removal unit 205.
The false detection check bit removal unit 205 removes, from the code word
transferred from the LDPC decode unit 202, the bits in the false detection check bit
insertion positions p notified from the false detection check bit control unit 203 (the
82nd, 84*, 86th, 88th, 90th, 92nd, 94th, 96th, 98th, 100th, 102nd, 104th, 106th, and 108th
bits). The false detection check bit removal unit 205 also removes the 324 redundant
bits from the code word, and outputs the remaining 310-bit sequence as the decoded
information bit sequence.
[0133]
Due to the above structure, the decoding device 200 can lower the rate of
false detection whereby a received code word that satisfies the parity check but is an
erroneous code word is judged as an error-free code word.
[0134]
As has been described above, the coding device pertaining to the present
Embodiment 1 inserts false detection check bits, which are known bits
predetermined by the transmitter device and the receiver device, into a code word
generated based on an information bit sequence to be transmitted. Specifically, in
the generated code word, the false detection check bits are inserted into portions that
would be different from another code word to be received and decoded by the
receiver device, said another code satisfying the parity check and having a close
Hamming distance with respect to the generated code word. In the receiver device, if
the false detection check bits inserted in the received and decoded code word are
different from the predetermined known bits, then the received and decoded code
word is judged to be different from the corresponding code word generated by the
transmitter device, and thus a failure in reception of an error-free code word can be
detected. In addition, the above structure allows lowering the rate of false detection
wherein at least part of a received and decoded code word that (i) is judged as
error-free by the parity check in the receiver device but (ii) is not a corresponding
code word generated by the transmitter device is judged as error-free.

The following describes differences between the present embodiment and
the technology of Patent Literature 1 that, as with the present embodiment, makes
use of known bits holding known values in error-correction decoding processing.
[0135]
Patent Literature 1 describes a decoding method utilizing the maximum a
posteriori probability decoding. According to this decoding method, when decoding
transmission words that include known bits with known values, a decoded word
candidate in which a known value in a part of a transmission word has been changed
to another value is excluded from the decoded word candidates.
[0136]
As a result of comparison between the present Embodiment 1 and the
technology of Patent Literature 1 in terms of the structure of a transmitter device, the
following difference was found: as the technology of Patent Literature 1 utilizes
known bits included in data to be transmitted, such as synchronization bytes of
MPEG-2TS packets, it neither inserts known bits into a data sequence, nor controls
insertion positions of the known bit. That is, the technology of Patent Literature 1
cannot achieve the effect of improvement in the error detection accuracy that can be
obtained by, as described in the present Embodiment 1, inserting known bits into bit
positions that hold different values between an error-free code word and a
minimum-distance code word.
[0137]
Moreover, as a result of comparison between the present Embodiment 1 and
the technology of Patent Literature 1 in terms of the structure of a receiver device,
the following difference was also found: according to the decoding method
disclosed in Patent Literature 1, decoding is performed after removing, from a
plurality of decoded word candidates, one or more decoded word candidates whose
known bits hold different values. That is to say, the technology of Patent Literature 1
does not relate to an error detection method. It seems like the technology of Patent
Literature 1 has a structure in which a false detection check is performed prior to
error-correction coding. However, the MAP decoding used in the technology of
Patent Literature 1 is a decoding method for maximizing a posteriori probability of
each bit of code words. Therefore, a code word sequence obtained as a result of the
MAP decoding is not necessarily composed of code words of the code used by the
transmitter device. Even if such a code word sequence were composed of code
words of the code used by the transmitter device, there is a possibility that values of
bits at positions of known bits change in the course of the MAP decoding. For the
above reasons, according to the technology of Patent Literature 1 the effect of
improvement in the error detection accuracy cannot be achieved by the false
detection check.
[0138]
The technology of the present Embodiment 1 for preventing a false
detection may incorporate any decoding method in error-correction decoding. Hence,
the false detection prevention technology of the present Embodiment 1 may be
adapted to the decoding method of Patent Literature 1. In which case, the effect of
improvement in the error detection accuracy, which cannot be achieved solely by the
decoding method of Patent Literature 1, can be achieved.
[Embodiment 2]
The above Embodiment 1 discloses technology for controlling insertion
positions of false detection check bits so as to lower the rate of false detection
whereby an erroneous code word is judged as an error-free code word. The present
Embodiment 2 provides a description of technology that makes use of information to
be transmitted rather than inserting predetermined bit values as false detection check
bits, and therefore can suppress reduction in the throughput of data transfer caused
by insertion of false detection check bits.

In a case where the information length of an actual information bit sequence
is shorter than the information length of an information bit portion determined by
the codeword length and code rate, a code word is generated in the following
manner: known bits that have been predetermined as pseudo information (also called
shortening bits) are appended so that the information length of the actual
information bit sequence equals the information length of the information bit portion.
In other words, appending of the known bits has been determined from the
beginning, and thus does not result in a throughput reduction.
[0139]
In view of the above, the present Embodiment 2 describes an example of
how to raise the rate of prevention of false detection whereby an erroneous code
word is judged as an error-free code word by the parity check. More specifically, the
rate of prevention of false detection can be raised by inserting known bits that have
been determined to be appended as false detection check bits into the information bit
sequence.
[0140]
Below, a specific example of a conventional method for inserting shortening
bits, as well as a concept of insertion of false detection check bits by using
shortening bits pertaining to the present Embodiment 2, will be explained in that
order with reference to FIGs. 12A-1 through 13E.
[0141]
As an LDPC code is a block code, coding of the LDPC code is performed as
follows. Coding processing is performed on the input, namely a K-bit information
bit sequence; an M-bit redundant bit sequence is appended to the K-bit information
bit sequence; and a resultant N-bit code word is output. FIGs. 12A-1 and 12B-1 each
illustrate an input information bit sequence. FIGs. 12A-2 and 12B-2 each illustrate a
state where the information bit sequence has been divided into smaller information
bit sequences each having the size for an LDPC code block. FIGs. 12A-3 and 12B-3
each illustrate a state where the divided information bit sequences have been coded.
FIGs. 12A-4 and 12B-4 each illustrate a code word to be ultimately output.
[0142]
Assume a case where the input information bit sequence is shorter than K
bits. In a communication system conforming to IEEE 802.1 In and the like, the size
of data to be transmitted varies depending on the time of transmission. Hence, in
many cases, the size of data to be transmitted is not an integral multiple of the
information bit length K of the LDPC code. In a case where the size of data to be
transmitted is not an integral multiple of the information bit length K of the LDPC
code, if K-bit information bit sequences are input to the LDPC coder one after
another, then the number of bits contained in the last LDPC code block to be coded
is less than K. At this time, as shown in FIG. 12A-3, a K-bit sequence is generated
by appending a certain number of (in this example, S) known bits (shortening bits,
usually having a value "0") at the end or in a given position of information bits of
the last LDPC code block, so that the last LDPC code block would have a length of
K bits. Thereafter, the generated K-bit sequence is coded. After the coding is
performed, the appended S known bits are removed, and the resultant sequence is
output as a code word. The above coding method is called shortening coding
(alternatively, a code word obtained without removing the S known bits may be
transmitted as-is).
Similarly, in a case where a plurality of blocks are to be coded after an
attempt to allocate the same number of information bits to each block, one of the
blocks ends up with containing less than K information bits as shown in FIG. 12B-2.
In this case also, the shortening coding is performed by inserting S known bits. As
indicated in Non-Patent Literature 2, in the case of IEEE 802.1 In, known bits are
appended to the end of an information bit sequence of each LDPC code block, so
that each LDPC code block contains an equal number of known bits.
[0143]
The known bits used for shortening coding (hereinafter referred to as
shortening bits) are appended in order to make a bit sequence input to the coder
contain K bits. The positions in which the shortening bits are appended in the input
sequence do not affect processing of shortening coding itself. An error-correction
coding device pertaining to the present embodiment inserts false detection check bits
into insertion positions of shortening bits—i.e., the error-correction coding device
uses the same known bits for the purposes of shortening and false detection check.
This helps lower the degree of throughput reduction.
[0144]
With reference to FIGs. 13A to 13E, the following describes a concept of
coding pertaining to the present Embodiment 2, which uses the shortening bits
shown in FIGs. 12A-3 and 12B-3. FIG. 13A shows an information bit sequence. As
this information bit sequence has a smaller size than an LDPC code block,
shortening bits are appended to this information bit sequence as shown in FIG. 13B.
[0145]
At this time, the transmitter device and the receiver device both use known
values for the shortening bits. In addition, the shortening bits may be inserted
anywhere in the information bit sequence. Thus, the appended shortening bits and
information bits in the information bit sequence are rearranged so as to relocate the
shortening bits to the insertion positions of false detection check bits of the
information bit sequence. Put another way, the shortening bits are inserted into the
insertion positions of the false detection check bits within the information bits.
[0146]
As a result, a pre-coding bit sequence shown in FIG. 13C is generated. Then,
redundant bits (parity bits) based on the LDPC code are appended to the generated
pre-coding bit sequence as shown in FIG. 13D. Thereafter, the inserted shortening
bits may be removed. Note that in FIGs. 13C and 13D, the rectangular portions with
diagonal lines represent the information bit sequence.
[0147]
Also note that FIG. 13E simply shows an example of a conventional code
word that has been coded with shortening bits appended to the end of information
bits.
[0148]
As set forth above, the present Embodiment 2 incorporates shortening bits
to be appended as false detection check bits.

FIG. 14 shows a functional structure of a coding device 1400 pertaining to
Embodiment 2. As shown in FIG. 14, the coding device 1400 includes an LDPC
coding parameter storage unit 101, an LDPC coding unit 104, a false detection
check bit control unit 1401, a shortening bit insertion unit 1402, a rearrangement
unit 1403, and a shortening bit removal unit 1404. Note that the coding device 1400
is different from the coding device 100 described in Embodiment 1 in including the
shortening bit insertion unit 1402, the rearrangement unit 1403, and the shortening
bit removal unit 1404. The following description of the coding device 1400 is given
with a focus on the differences between the coding device 1400 and the coding
device 100 described in Embodiment 1.
[0149]
The false detection check bit control unit 1401 notifies information
indicating the number D of false detection check bits and false detection check bit
insertion positions p, which is transferred from the LDPC coding parameter storage
unit 101, to the rearrangement unit 1403.
[0150]
The shortening bit insertion unit 1402 inserts shortening bits to the end of
the information bit sequence. Here, the number of shortening bits to be inserted is
provided to the shortening bit insertion unit 1402 as control information. If the
transmitter device and the receiver device both acknowledge where the shortening
bits are inserted, then the insertion positions of the shortening bits are not limited to
the end of the information bit sequence. Also, the shortening bits are inserted so that
the information length of the information bit sequence is a multiple of the
information length of an information bit sequence portion of the code word to be
generated. In other words, the number of shortening bits to be inserted is determined
based on the information length of the information bit sequence and the information
length of the information bit sequence portion of the code word to be generated.
[0151]
The rearrangement unit 1403 rearranges the order of bits in the information
bit sequence into which the shortening bits have been inserted. More specifically,
this rearrangement is done by relocating the shortening bits currently positioned to
the end of the information bit sequence to the false detection check bit insertion
positions.
[0152]
The shortening bit removal unit 1404 removes the shortening bits included
in the code word bits output from the LDPC coding unit 104. The shortening bit
removal unit 1404 outputs the code word bits from which the shortening bits have
been removed.
[0153]
FIG. 15 shows a functional structure of a decoding device 1500 pertaining
to Embodiment 2. As shown in FIG. 15, the decoding device 1500 includes an
LDPC coding parameter storage unit 201, an LDPC decode unit 202, a false
detection check bit control unit 203, an error detection unit 204, a shortening bit
insertion unit 1501, a rearrangement unit 1502, a rearrangement unit 1503, and a
shortening bit removal unit 1504. As with the above description of the coding device
1400, the following description of the decoding device 1500 is given with a focus on
the differences between the decoding device 1500 and the decoding device 200
described in Embodiment 1.
[0154]
The shortening bit insertion unit 1501 inserts shortening bit likelihoods to
the end of the received code word. It should be noted here that the shortening bit
likelihoods are values corresponding to the shortening bits inserted by the
transmitter device and to a decoding method used in the LDPC decode unit 202. For
example, in a case where the transmitter device inserts "0" as shortening bits and the
LDPC decode unit 202 performs soft-decision decoding, a positive number having a
large absolute value should be inserted as each shortening bit likelihood. On the
other hand, in a case where the LDPC decode unit 202 performs hard-decision
decoding, "0" is inserted as each shortening bit likelihood. The shortening bit
insertion unit 1501 transfers the received code word into which the shortening bit
likelihoods have been inserted to the rearrangement unit 1502.
[0155]
The rearrangement unit 1502 rearranges bits in the received code word
transferred from the shortening bit insertion unit 1501. More specifically, this
arrangement is done by relocating the shortening bit likelihoods inserted by the
shortening bit insertion unit 1501 to the positions designated by the false detection
check bit positions p transferred from the false detection check bit control unit 203.
The rearrangement unit 1502 transfers the received code word with rearranged bits
to the LDPC decode unit 202. The shortening bits, which were deleted at the time of
transmission, are restored. Thereafter, the shortening bits can be rearranged so as to
return back to the positions where they were originally inserted as false detection
check bits.
[0156]
The rearrangement unit 1503 rearranges bits in the decoded code word
obtained as a result of the decode processing performed by the LDPC decode unit
202. More specifically, this rearrangement is done by relocating the shortening bits
at the shortening bit insertion positions p to the end of the decoded code word bits.
The rearrangement unit 1503 transfers the decoded code word with rearranged bits
to the shortening bit removal unit 1504.
[0157]
The shortening bit removal unit 1504 removes the shortening bits and parity
bits from the end of the decoded code word bits that have been rearranged and
transferred from the rearrangement unit 1503. Consequently, the shortening bit
removal unit 1504 obtains a J-bit information bit sequence. The shortening bit
removal unit 1504 outputs the J-bit information bit sequence as the output from the
decoding device 1500.
[0158]
This concludes the description of functional structures of the coding device
1400 and decoding device 1500 pertaining to Embodiment 2.

The following describes coding operations performed by the coding device
1400 with reference to the flowchart of FIG. 16. Note that part of the flowchart of
FIG. 16 that is the same as the flowchart of FIG. 9 pertaining to Embodiment 1 is
omitted from the following description.
[0159]
As shown in FIG. 16, the false detection check bit control unit 1401 notifies
the rearrangement unit 1403 of the insertion positions and the number of false
detection check bits (step SI601).
[0160]
The shortening bit insertion unit 1402 in the coding device 1400 receives
control information indicating the number of shortening bits to be inserted, appends
the number of shortening bits indicated by the control information to the end of the
input information bit sequence, and outputs the resultant information bit sequence to
the rearrangement unit 1403 (step SI602).
[0161]
Next, the rearrangement unit 1403 rearranges bits in the received
information bit sequence to which the shortening bits have been appended. More
specifically, the rearrangement unit 1403 performs this rearrangement so that, per an
information length of an information bit portion in a generated code word, the
shortening bits are located in the false detection check bit insertion positions notified
from the false detection check bit control unit 1401. Thereafter, the rearrangement
unit 1403 transfers the rearranged pre-coding bit sequence to the LDPC coding unit
104 (step SI603).
[0162]
The LDPC coding unit 104 generates a code word by coding the transferred
pre-coding bit sequence in accordance with the parity check matrix notified from the
LDPC coding parameter storage unit, and transfers the generated code word to the
shortening bit removal unit 1404 (step S905).
[0163]
The shortening bit removal unit 1404 removes shortening bits from the
generated code word (step SI604), and ends the processing.
[0164]
The above has explained the flowchart of FIG. 16 in relation to the case
where shortening bits have been inserted. However, when shortening bits need not
be inserted into an information bit sequence—i.e., when the information length of
the information bit sequence equals the information length of a block pertaining to a
coding method utilizing quasi-cyclic codes, false detection check bits holding
known values shall be inserted into specified positions as described in the above
Embodiment 1.
[0165]
Described below is one specific example of this coding method. First, a
description is given of a specific example where the number of shortening bits to be
appended is 14 or more.
[0166]
To be more specific, the following description relates to the case of J = 300,
as one exemplary case where the sequence length J of the input information bit
sequence is shorter than the information block length K (324) by at least the number
D of false detection check bits (14). At this time, the number S of shortening bits is
24. The coding mode information is input to the LDPC coding parameter storage
unit 101. In accordance with the input coding mode information, the LDPC coding
parameter storage unit 101 sets parameters required for LDPC coding in the false
detection check bit control unit 102 and the LDPC coding unit 104. Examples of the
parameters set in the false detection check bit control unit 102 include the number of
false detection check bits, namely 14, and false detection check bit insertion
positions p (the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th, 100th, 102nd, 104th,
106th, and 108th bits). Examples of the parameters set in the LDPC coding unit 104
include the code length N (648 bits), the information bit length K (324 bits), the
redundant bit length M (324 bits), the parity check matrix H (or an address of the
memory storing therein information relating to the parity check matrix H), and in a
case where different processing is performed for each coding mode, a signal
indicating which processing is to be performed.
[0167]
The input information bit sequence composed of 300 bits is input to the
shortening bit insertion unit 1402. The shortening bit insertion unit 1402 inserts 24
shortening bits into the end of the input information bit sequence. The shortening bit
insertion unit 1402 transfers the information bit sequence, now having a 324-bit
length as a result of insertion of the shortening bits, to the rearrangement unit 1403.
[0168]
The rearrangement unit 1403 rearranges the 324-bit information bit
sequence. To be more specific, the rearrangement unit 1403 rearranges (i) the
shortening bits, which make up the last 14 bits of the information bit sequence, and
(ii) the bits located in the false detection check bit insertion positions p, namely the
82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th, 100th, 102nd, 104th, 106th, and 108th bits.
This makes all the bits located in the false detection check bit insertion positions p
the known bits. The rearrangement unit 1403 transfers the rearranged information bit
sequence to the LDPC coding unit 104. In performing the rearrangement processing,
the rearrangement unit 1403 may use any rearrangement rule as long as it can
rearrange the bits located in the false detection check bit insertion positions p and
the shortening bits.
[0169]
The LDPC coding unit 104 performs processing of LDPC coding on the
input pre-coding information bit sequence, generates a 324-bit redundant bit
sequence, combines the generated 324-bit redundant bit sequence with the 324-bit
coded information bit sequence, and outputs a result of the combination as a 648-bit
code word to the shortening bit removal unit 1404.
[0170]
The shortening bit removal unit 1404 performs processing of removing S
(24) shortening bits inserted by the shortening bit insertion unit 1402 from the code
word bits. As a result of removal of the shortening bits, the code word has a 624-bit
length (300 + 324 = 624). The shortening bit removal unit 1404 outputs the 624-bit
code word as the output from the coding device 1400.
[0171]
A code word is generated through the above procedure.
[0172]
Meanwhile, as described in the above Embodiment 1, at least 14 bits need to
be inserted as false detection check bits in order to perform error detection for all of
the minimum-distance code words. However, in some cases, the number of
shortening bits is smaller than 14 depending on the information length of the
information bit sequence. In consideration of such cases, the following describes a
specific example where the number of shortening bits to be appended is smaller than
14. It should be noted that the coding device itself merely appends a specified
number of shortening bits and rearrange the shortening bits by relocating them to the
specified false detection check bit insertion positions. That is to say, the coding
device basically operates in accordance with the flowchart of FIG. 16. Generality of
the flowchart of FIG. 16 is retained regardless of the number of shortening bits to be
appended.
[0173]
Next, a description is given of a case where the number S of shortening bits
is smaller than the number D of false detection check bits. As one example, the
following discusses the case of J = 320 bits. More specifically, in this case, S = 4
andD=14.
[0174]
An input information bit sequence composed of 320 bits is input to the
shortening bit insertion unit 1402. The shortening bit insertion unit 1402 inserts the
4 shortening bits into the end of the input information bit sequence. The shortening
bit insertion unit 1402 transfers the information bit sequence, now having a 324-bit
length as a result of insertion of the shortening bits, to the rearrangement unit 1403.
[0175]
The rearrangement unit 1403 rearranges the 324-bit information bit
sequence. More specifically, the rearrangement unit 1403 rearranges (i) the
shortening bits, which make up the last four bits of the information bit sequence, and
(ii) the four bits located in the false detection check bit insertion positions p, namely,
the 82nd, 84th, 86th, and 88th bits, for example. This makes the 4 bits located in the
false detection check bit insertion positions p the known bits. The rearrangement
unit 1403 transfers the rearranged information bit sequence to the LDPC coding unit
104. In the rearrangement processing performed by the rearrangement unit 1403, the
four positions to be selected from the false detection check bit insertion positions p
are not limited to the above-described four positions, as long as they are known to
the transmitter device and the receiver device. Selecting four positions different
from the above-described four positions merely changes a code word on which the
false detection is performed, and does not change the rate at which the false
detection can be made on average. In the above case, however, not all of the
minimum-distance code words can be detected because the number S of shortening
bits is smaller than the number D of false detection check bits. Thus, in the above
case, the rate at which the false detection can be made is low as compared to the
case of S > D.
[0176]
The LDPC coding unit 104 performs processing of LDPC coding on the
input pre-coding information bit sequence, generates a 324-bit redundant bit
sequence, combines the generated 324-bit redundant bit sequence with the 324-bit
coded information bit sequence, and outputs a result of the combination as a 648-bit
code word to the shortening bit removal unit 1404.
[0177]
The shortening bit removal unit 1404 performs processing of removing the
S (4) shortening bits, which have been inserted by the shortening bit insertion unit
1402, from the code word bits. As a result of removal of the shortening bits, the
code word has a 644-bit length (320 + 324 = 644). The shortening bit removal unit
1404 outputs the 644-bit code word as the output from the coding device 1400.
[0178]
In the above description, the coding device 1400 includes the shortening bit
insertion unit 1402 and performs insertion of the shortening bits on its own.
However, the present embodiment is not limited to this structure. Alternatively, for
example, an information bit sequence into which shortening bits have already been
inserted may be input to the coding device 1400. In this case, the coding device
1400 need not include the shortening bit insertion unit 1402 because it does not
perform the processing of inserting the shortening bits on its own.
[0179]
With reference to the flowchart of FIG. 17, the following describes
operations performed by the decoding device 1500 upon receiving the code word
that has been generated in the above manner, modulated, and transmitted. In the
following description, part of the flowchart of FIG. 17 that is the same as the
flowchart of FIG. 11 pertaining to Embodiment 1 is omitted.
[0180]
In step SI701, the false detection check bit control unit 203 notifies the
error detection unit 204, the rearrangement unit 1502, and the shortening bit removal
unit 1504 of false detection check bit insertion positions, as well as known values
used as the false detection check bits.
[0181]
Next, the shortening bit insertion unit 1501 inserts shortening bit likelihoods
into the received code word in accordance with the control information (step SI702).
At this time, the number of the shortening bit likelihoods to be inserted corresponds
to the number of shortening bits and the number of parity bits generated based on
the shortening bits.
[0182]
Then, the rearrangement unit 1502 rearranges bits in the received code word
to which the shortening bit likelihoods have been appended, by relocating the
shortening bit likelihoods to the false detection check bit insertion positions (step
SI 703).
[0183]
In step SI704, the rearrangement unit 1503 extracts the bits located at the
false detection check bit insertion positions and relocates the extracted bits to the
end of the information bit sequence, so as to place the shortening bit likelihoods
inserted in the decoded code word in the end of the information bit sequence.
[0184]
Thereafter, the shortening bit removal unit 1504 removes the parity bits as
well as the shortening bits relocated to the end from the rearranged decoded code
word (step SI705), outputs the decoded information bit sequence, and ends the
processing.
[0185]
The following describes a specific example of operations to be executed
when the decoding device performs decoding in accordance with the flowchart of
FIG. 17.
[0186]
More specifically, the following description relates to operations performed
by the decoding device 1500 with use of an LDPC code defined by the parity check
matrix Hb648 shown in FIG. 1 and having a code length of 648 bits and a code rate
of 1/2 as specified in IEEE 802.1 In.
[0187]
Before the received code word is input, the coding mode information and
the number of shortening bits are input to the decoding device 1500. In accordance
with the coding mode information, the LDPC coding parameter storage unit 201
transfers LDPC coding parameters which are stored in correspondence with the
received coding mode information to the LDPC decode unit 202 and the false
detection check bit control unit 203. Examples of the parameters set in the false
detection check bit control unit 203 include the number of false detection check bits,
namely 14, and false detection check bit insertion positions p (the 82nd, 84th, 86th,
88th, 90th, 92nd, 94th, 96th, 98th, 100th, 102nd, 104th, 106th, and 108th bits). Examples of
the parameters set in the LDPC decode unit 202 include the code length N (648 bits),
the information bit length K (324 bits), the redundant bit length M (324 bits), the
parity check matrix H (or an address of the memory storing therein information
relating to the parity check matrix H), and in a case where different processing is
performed for each coding mode, a signal indicating which processing is to be
performed.
[0188]
The following description pertains to a case where the number J of bits in
the information bit sequence is 300, and the number S of shortening bits is 24. After
the parameters to be used are set, the LDPC decode processing begins for the
received code word. The received 624-bit code word is input to the shortening bit
insertion unit 1501. The shortening bit insertion unit 1501 inserts 24-bit shortening
bit likelihoods between an information bit portion and a parity bit portion (i.e., to the
end of the information bit portion) in the received code word. As a result, the
information bit portion in the received code word increases by 24 bits. The
shortening bit insertion unit 1501 transfers the received code word into which the
shortening bit likelihoods have been inserted to the rearrangement unit 1502.
[0189]
The rearrangement unit 1502 rearranges bits in the received code word.
Since the number D of false detection check bits is 14, the rearrangement unit 1320
performs the rearrangement by inserting the last 14 likelihoods of the information bit
portion in the received code word (i.e., the bits corresponding to the inserted 24-bit
shortening bit likelihoods) into the positions specified by the false detection check
bit insertion positions p (the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th, 100th,
102nd, 104th, 106th, and 108th bits). The rearrangement unit 1502 transfers the
received code word with rearranged bits to the LDPC decode unit 202.
[0190]
The LDPC decode unit 202 performs LDPC decode processing on the
received code word, and transfers the resultant code word to the error detection unit
204 and the rearrangement unit 1503.
[0191]
The rearrangement unit 1503 relocates the bits located at the false detection
check bit insertion positions p (the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th,
100th 102nd, 104th, 106th, and 108th bits) to the end of the information bit portion of
the decoded code word. The rearrangement unit 1503 transfers the decoded code
word with rearranged bits to the shortening bit removal unit 1504.
[0192]
The shortening bit removal unit 1504 removes the last 24 bits of the
information bit portion, as well as 324 bits corresponding to the parity bit sequence,
from the decoded code word transferred from the rearrangement unit 1503. The
shortening bit removal unit 1504 outputs the 300 bits obtained as a result of the
above removal as a decoded information bit sequence.
[0193]
The structure of the error detection unit 204 and the processing performed
thereby (i.e., processing of comparing the shortening bit likelihoods inserted as false
detection check bits with known values used as shortening bits) are the same as
those described in Embodiment 1. The descriptions of such structure and processing
are hence omitted.
[0194]
The following description pertains to a case where the number J of
information bits is 320 and the number S of shortening bits is 4. The received
620-bit code word is input to the shortening bit insertion unit 1501. The shortening
bit insertion unit 1501 inserts 4-bit shortening bit likelihoods between an
information bit portion and a parity bit portion (i.e., to the end of the information bit
portion)in the received code word. As a result, the information bit portion of the
received code word increases by 4 bits. The shortening bit insertion unit 1501
transfers the received code word into which the shortening bit likelihoods have been
inserted to the rearrangement unit 1502.
[0195]
The rearrangement unit 1502 rearranges bits in the received code word. To
be more specific, the rearrangement unit 1502 performs the rearrangement by
inserting the last 4 likelihoods of the information bit portion in the received code
word (i.e., the bits corresponding to the inserted 4-bit shortening bit likelihoods) into,
among the false detection check bit insertion positions p, the positions to which the
shortening bits have been inserted by the rearrangement unit 1403 of the coding
device 1400 in the transmitter device (in the above example, the 82nd, 84th, 86th and
88 bits). The rearrangement unit 1502 transfers the received code word with
rearranged bits to the LDPC decode unit 202.
[0196]
The LDPC decode unit 202 performs LDPC decode processing on the
received code word, and transfers the resultant code word to the error detection unit
204 and the rearrangement unit 1503.
[0197]
The rearrangement unit 1503 relocates, out of the bits located in the false
detection check bit insertion positions p, the bits located in the four positions into
which the shortening bits have been inserted (the 82nd, 84th, 86th and 88th bits) to the
end of the information bit portion of the decoded code word. The rearrangement unit
1503 transfers the decoded code word with rearranged bits to the shortening bit
removal unit 1504.
[0198]
The shortening bit removal unit 1504 removes the last 4 bits of the
information bit portion, as well as 324 bits corresponding to the parity bit sequence,
from the decoded code word transferred from the rearrangement unit 1503. The
shortening bit removal unit 1504 outputs the 320 bits obtained as a result of the
above removal as a decoded information bit sequence.
[0199]
In Embodiment 1, insertion of false detection check bits reduces the
throughput of data transfer in proportion to the inserted false detection check bits. In
contrast, as has been described in the present Embodiment 2, such a decrease in the
throughput can be prevented by following structure: in a case where the length of the
information bit sequence is not a multiple of the length of the information bit portion
calculated based on the code length and code rate of a code word, known bits that
have been originally determined to be inserted to supplement shortage in bits are
used as false detection check bits. By using something that has been originally
determined to be inserted, there is no necessity to insert new false detection check
bits. As a result, a throughput reduction caused by insertion of false detection check
bits can be prevented.
[0200]
With the above-described structure, the decoding device 1500 can decode
the received code word and perform error detection on the result of the decoding.
The decoding device 1500 pertaining to the present embodiment is configured not
only to perform a parity check with use of the parity check matrix H of the LDPC
code, but also to check whether a code word obtained after the LDPC decoding is a
minimum-distance code word in relation to a corresponding error-free code word.
Accordingly, the decoding device 1500 pertaining to the present embodiment can
increase the error detection accuracy to a significant extent as compared to a
conventional error detection that only uses the parity check.
[0201]
The present embodiment has described the structure in which shortening
bits are substituted for false detection check bits, and other known bits are not used.
However, the present invention is not limited to such structure. For example, when
the number of shortening bits is smaller than the number of false detection check
bits, known bits may be inserted in addition to the shortening bits so that the total
number of known bits and shortening bits equals the number of false detection check
bits. This way, the false detection accuracy can still be improved even when the
number of shortening bits is smaller than the number of false detection check bits.
[Embodiment 3]
The above embodiment 2 has described the exemplary case where known
bits are inserted as false detection check bits when there is shortage in the
information length. The present Embodiment 3 describes an exemplary case where
other bit values are used as false detection check bits.

When the information length of an information bit portion of a code word,
which is determined by a codeword length and a code rate, is shorter than the
information length of an information bit sequence to be transmitted, a plurality of
code words are generated to transmit data pieces of the information bit sequence to
be transmitted. At this time, each of the generated code words is transmitted with
block ID bits appended thereto, the block ID bits indicating the order in which the
code word is located among all the code words carrying the information bit
sequence to be transmitted. As shown in FIG. 18A, each of the code words
generated in this case is composed of the following bits: (i) block ID bits that make
up the first B bits; (ii) J information bits; and (iii) M parity bits (redundant bits).
Note that K + M bits represent the codeword length, and K bits and M bits are
determined by the code rate.
[0202]
In the present Embodiment 3, values of bits of the appended block ID are
inserted as false detection check bits. The transmitted data pieces are basically
received in the order indicated by their respective block IDs, and therefore should be
decoded according to that order. When the receiver device restores a block ID by
extracting bits from the positions where the false detection check bits have been
inserted, in some cases the number indicated by that block ID may be different from
the order in which the corresponding data piece was received. In such cases, it is
assumed that the transmitted data pieces have changed during data transfer, and
therefore the received code words are judged as erroneous code words, even if they
were judged as error-free code words by the QC-LDPC parity check.
[0203]
A description is now given of block ID bits included in an LDPC code
block. In order to carry out a block retransmission method, it is necessary for the
transmitter device and the receiver device to acknowledge which blocks have been
or have not been properly transmitted, and which blocks should be selectively
retransmitted. One example way to acknowledge such blocks is to insert block ED
bits, which are block identifiers, into the LDPC code block. This method makes use
of part of bits included in the LDPC code block as block ID bits.
[0204]
FIGs. 18A through 18D show a bit sequence of an LDPC code into which
block ID bits have been inserted. As shown in FIG. 18A, in a bit sequence of an
LDPC code into which block ID bits have been inserted, a K-bit information bit
sequence of the LDPC code block is composed of B block ID bits and J information
bits. Furthermore, code word bits obtained by LDPC coding are made up of (i) the
K-bit information sequence including the block ID bits (K = B + J), and (ii) an M-bit
parity bit sequence appended to the K-bit information sequence. In addition, values
inserted in the block ID bits differ in each LDPC code block.
[0205]
Note that the block ID bits are inserted to identify the corresponding LDPC
code block. Accordingly, there are no rules regarding where the block ID bits should
be inserted, as long as the transmitter device and the receiver device both
acknowledge the insertion positions of the block ID bits. In view of the above, an
error-correction coding device pertaining to the present Embodiment 3 inserts block
ID bits into insertion positions of false detection check bits—i.e., the
error-correction coding device uses block ID bits as known bits both for
identification of the corresponding LDPC code block and for a false detection check.
This helps lower the degree of throughput reduction.
[0206]
FIG. 18A shows a code word to which conventional block ID bits have been
appended. In contrast, in the present Embodiment 3, block ID bits appended to the
information bit sequence are inserted into the false detection check bit insertion
positions that are counted from the beginning of the information bit portion as
shown in FIG. 18B. As a result, a pre-coding bit sequence shown in FIG. 18C is
generated. And finally, a code word shown in FIG. 18D is generated by coding the
pre-coding bit sequence. Note that in FIGs. 18C and 18D, the rectangular portions
with diagonal lines represent the information bit portion, and each blank rectangular
portion sandwiched between two rectangular portions with diagonal lines represents
a block ID bit inserted as a false detection check bit.

FIG. 19 shows a functional structure of a coding device 1900 pertaining to
Embodiment 3. As shown in FIG. 19, the coding device 1900 includes an LDPC
coding parameter storage unit 101, an LDPC coding unit 104, a false detection
check bit control unit 1901, and a rearrangement unit 1902.
[0207]
The false detection check bit control unit 1901 notifies the rearrangement
unit 1902 of the false detection check bit insertion positions notified from the LDPC
coding parameter storage unit 101.
[0208]
The rearrangement unit 1902 generates a pre-coding bit sequence by
rearranging the information bit sequence so as to relocate the information bits that
make up a block ID to the false detection check bit insertion positions notified from
the false detection check bit control unit 1901. Then, the rearrangement unit 1902
outputs the generated pre-coding bit sequence to the LDPC coding unit 104.
[0209]
FIG. 20 shows a functional structure of a decoding device 2000 pertaining
to Embodiment 3. As shown in FIG. 20, the decoding device 2000 includes an
LDPC coding parameter storage unit 201, an LDPC decode unit 202, a false
detection check bit control unit 203, a parity bit removal unit 2001, a rearrangement
unit 2002, and an error detection unit 2003.
[0210]
The parity bit removal unit 2001 removes M parity bits from the decoded
code word bits output from the LDPC decode unit 202. Thereafter, the parity bit
removal unit 2001 transfers only the information bit sequence (i.e., the decoded code
word bits excluding the removed parity bits) to the rearrangement unit 2002.
[0211]
The rearrangement unit 2002 extracts the bits located at the false detection
check bit insertion positions, and relocates the extracted bits to the front of the
information bit sequence. The rearrangement unit 2002 outputs a decoded
information bit sequence, which is obtained by relocating the bits in the false
detection check bit insertion positions to the front of the information bit sequence.
[0212]
The error detection unit 2003 performs error detection on the decoded code
word bits transferred from the LDPC decode unit 202. The error detection unit 2003
outputs the error detection result as the output from the decoding device 2000.
[0213]
FIG. 21 shows a detailed structure of the error detection unit 2003. As
shown in FIG. 21, the error detection unit 2003 includes a parity check subunit 211
and a false detection check subunit 2101. The parity check subunit 211 performs the
same processing as the parity check subunit 211 described in Embodiment 1. Hence,
in the following description, a detailed explanation on the parity check subunit 211
is omitted.
[0214]
When the result of the parity check shows "0", the false detection check
subunit 2101 checks the block ID bits inserted in the false detection check bit
insertion positions pi. The decoded code word cd, the result of the parity check, and
the control information indicating the number, values and insertion positions of false
detection check bits are input to the false detection check subunit 2101. The false
detection check subunit 2101 judges whether any of the block ID bits holds an
erroneous value. If the result of the judgment is affirmative, then the false detection
check subunit 2101 judges the block ID bits as containing an error. When the block
ID bits inserted in the false detection check bit insertion positions pi contain an
error, its code word is a minimum-distance code word in relation to a corresponding
error-free code word, and accordingly the error detection check subunit 2101 judges
that there has been a false detection. On the other hand, if none of the block ID bits
holds an erroneous value, then the false detection check subunit 2101 judges that
there has been no false detection.
[0215]
One example of a case where a block ID bit holds an erroneous value is
when the value of the block ID bit does not match any of multiple values
predetermined by the transmitter device and the receiver device. For instance,
assume that the transmitter device and the receiver device have determined multiple
values to be used as block IDs in advance. Here, if the receiver device finds that at
least one of values held by the block ID bits obtained by extracting the bits located
in the false detection check bit insertion positions does not match any of the
determined multiple values, then the stated at least one of values is judged as an
erroneous value.
[0216]
Another example of a case where a block ID bit holds an erroneous value is
when the block ID bit is found as not following the rule predetermined by the
transmitter device and the receiver device as a result of comparing the value of the
block ID bit with values of the previous and next block ID bits. For instance, assume
that LDPC code blocks included in one PHY frame are assigned block IDs
representing sequential numbers in order. Here, if a block ID that is not continuous
with a block ID of the previous or next LDPC code block is confirmed as a result of
the false detection check, then that block ID is judged as erroneous. To be more
specific, assume that the LDPC code blocks generated by the transmitter device are
assigned sequential numbers (e.g., 01, 02, 03, 04, and so on), and these sequential
numbers are also used as block IDs. Here, if the receiver device finds that the block
IDs located in the false detection check bit insertion positions show 01, 02, 13 and
04, respectively, then the LDPC code block including the block ID bits ^lKiwing""|!fi4-/!*-;:';;1
can be judged as erroneous, even when a corresponding code word is a code word of
the LDPC code.
[0217]
The false detection check subunit 2101 outputs a result of the false detection
check in format that can be interpreted by software or hardware. For example, the
false detection check subunit 2101 outputs "0" when no false detection has been
confirmed, and "1" when a false detection has been confirmed. The result of the
false detection check output from the false detection check subunit 2101 is also a
result of the error detection output from the error detection unit 2003. On the other
hand, when the result of the parity check shows "1", it is obvious that the decoded
code word cd is not a code word of the LDPC code and is thus erroneous. In this case,
there is no need to perform the false detection check. Therefore, in this case, the
false detection check subunit 2101 outputs "1" as a result of the false detection
check without performing the false detection check.

A description is now given of coding operations performed by the coding
device 1900 with reference to the flowchart of FIG. 22. In the following description,
only the operations that are different from those shown in FIG. 9 pertaining to
Embodiment 1 are discussed.
[0218]
The false detection check bit control unit 1901 notifies the rearrangement
unit 1902 of the false detection check bit insertion positions (step S2201).
[0219]
Then, the rearrangement unit 1902 generates a pre-coding bit sequence by
rearranging the information bit sequence in accordance with the notified false
detection check bit insertion positions, so as to relocate block ID bits, which are
originally located at the front of an information portion of the information bit
sequence, to the false detection check bit insertion positions. Thereafter, the
rearrangement unit 1902 transfers the generated pre-coding bit sequence to the
LDPC coding unit 104 (step S2202).
[0220]
Described below is one specific example of the coding operations
performed by the coding device 1900.
[0221]
More specifically, the following describes operations performed by the
coding device 1900 with use of an LDPC code defined by the parity check matrix
Hb648 shown in FIG. 1 and having a code length of 648 bits and a code rate of 1/2
as specified in IEEE 802.1 In.
[0222]
The transmitter device inputs, to the coding device 1900, coding mode
information as well as an information bit sequence containing block ID bits. Here,
assume that the first 8 bits of the information bit sequence are the block ID bits. In
this case, B = 8. The coding mode information is input to the LDPC coding
parameter storage unit 101. In accordance with the input coding mode information,
the LDPC coding parameter storage unit 101 sets parameters required for LDPC
coding in the false detection check bit control unit 1901 and the LDPC coding unit
104. Examples of the parameters set in the false detection check bit control unit
1901 include B false detection check bit insertion positions pi (e.g., the 82nd, 84th,
86th, 88th, 90th, 92nd, 94th, and 96th bits). Examples of the parameters set in the LDPC
coding unit 104 include the code length N (648 bits), the information bit length K
(324 bits), the redundant bit length M (324 bits), the parity check matrix H (or an
address of the memory storing therein information relating to the parity check matrix
H), and in a case where different processing is performed for each coding mode, a
signal indicating which processing is to be performed.
[0223]
The 324-bit information bit sequence containing the 8 block ID bits is input
to the rearrangement unit 1902.
[0224]
The rearrangement unit 1902 rearranges the 324-bit information bit
sequence. To be more specific, the rearrangement unit 1902 rearranges (i) the block
ID bits, which make up the first 8 bits of the information bit sequence, and (ii) the
bits located in the false detection check bit insertion positions pi, namely the 82nd,
84th, 86th, 88th, 90th, 92nd, 94th, and 96th bits. The makes the bits located in the false
detection check bit insertion positions pi the block ID bits. The rearrangement unit
1902 transfers the rearranged information bit sequence to the LDPC coding unit 104.
In performing the rearrangement processing, the rearrangement unit 1902 may use
any rearrangement rule as long as it can rearrange the bits located in the false
detection check bit insertion positions pi and the block ID bits.
[0225]
The LDPC coding unit 104 performs LDPC coding processing on the input
pre-coding information bit sequence, generates a 324-bit redundant bit sequence,
combines the generated 324-bit redundant bit sequence with the 324-bit coded
information bit sequence, and finally outputs a result of the combination as a 648-bit
code word as the output from the coding device 1900.
[0226]
With reference to the flowchart of FIG. 23, the following describes
operations performed by the decoding device 2000 upon receiving the code word
that has been generated in the above manner, modulated, and transmitted. In the
following description, only the operations that are different from those shown in FIG.
11 pertaining to Embodiment 1 are discussed.
[0227]
In step S2301, the error detection unit 2003 performs the error detection in
the following manner. After the parity check is performed by the parity check
subunit 211, the false detection check subunit 2101 judges whether the block ID bits
located in the false detection check bit insertion positions have the same values as
predicted bit values by comparing the block ID bits with the predicted bit values.
The false detection check subunit 2101 then outputs a result of the error detection.
For example, in a case where the transmitter device and the receiver device have
made it a rule to use some of predetermined values as block ID bits, the false
detection check subunit 2101 detects whether the block ID bits hold said some of
predetermined values. Alternatively, in a case where the order of sequential code
blocks is to be checked, the false detection check subunit 2101 detects whether each
block ID bit is continuous with a block ID bit of the previous or next LDPC code
block. In the latter case, the false detection check subunit 2101 detects whether each
block ID matches a block ID predicted from the block ID of the previous or next
LDPC code block by comparing the two (step S2301).
[0228]
The parity bit removal unit 2301 removes M parity bits from the decoded
code word, and transfers only the information bit sequence to the rearrangement unit
2002 (step S2302).
[0229]
Thereafter, the rearrangement unit 2002 rearranges the information bit
sequence by relocating the bits located in the false detection check bit insertion
positions to the front of the information bit sequence. The rearrangement unit 2002
generates a decoded information bit sequence through this procedure, and outputs
the decoded information bit sequence (step S2303).
[0230]
Described blow is one specific example of the decoding operations
performed by the decoding device 2000.
[0231]
More specifically, the following describes operations performed by the
decoding device 2000 with use of an LDPC code defined by the parity check matrix
Hb648 shown in FIG. 1 and having a code length of 648 bits and a code rate of 1/2
as specified in IEEE 802.1 In.
[0232]
Before the received code word is input, the coding mode information is
input to the decoding device 2000. In accordance with the coding mode information,
the LDPC coding parameter storage unit 201 transfers LDPC coding parameters
which are stored in correspondence with the received coding mode information to
the LDPC decode unit 202 and the false detection check bit control unit 203.
Examples of the parameters set in the false detection check bit control unit 203
include the number of false detection check bits, namely 14, and false detection
check bit insertion positions p (the 82nd, 84th, 86th, 88th, 90th, 92nd, 94th, 96th, 98th,
100th, 102nd, 104th, 106th, and 108th bits). Examples of the parameters set in the
LDPC decode unit 202 include the code length N (648 bits), the information bit
length K (324 bits), the redundant bit length M (324 bits), the parity check matrix H
(or an address of the memory storing therein information relating to the parity check
matrix H), and in a case where different processing is performed for each coding
mode, a signal indicating which processing is to be performed.
[0233]
After the parameters to be used are set, the decoding device 2000 begins
LDPC decode processing on the received code word. The received 648-bit code
word is input to the LDPC decode unit 202. At this time, in the received code word,
one bit may carry one-bit information (hard-decision decoding), or one bit may carry
multiple-bit information (soft-decision decoding). In the case of soft-decision
decoding, log-likelihood ratios of quantized code word bits may be used as the
received code word.
[0234]
The LDPC decode unit 202 performs LDPC decode processing on the
received code word, and transfers the resultant code word to the parity bit removal
unit 2001 and the error detection unit 2003.
[0235]
The parity bit removal unit 2001 removes M parity bits from the code word
transferred from the LDPC decode unit 202, and outputs only the K-bit information
bit sequence to the rearrangement unit 2002.
[0236]
The parity check subunit 211 included in the error detection unit 2003
checks whether the decoded code word Cj that has been input and the parity check
matrix H of the LDPC code that has been provided as control information satisfy the
parity check Hcd = 0. The parity check subunit 211 outputs a result of the parity
check in format that can be interpreted by software or hardware.
[0237]
When the result of the parity check shows "0", the false detection check
subunit 2101 checks the block ID bits inserted in the false detection check bit
insertion positions pi. The false detection check subunit 2101 judges whether any of
the block ID bits holds an erroneous value. If the result of the judgment is
affirmative, then the false detection check subunit 2101 judges the block ID bits as
containing an error. When the block ID bits inserted in the false detection check bit
insertion positions pi contain an error, its code word is a minimum-distance code
word in relation to a corresponding error-free code word, and accordingly the error
detection check subunit 2101 judges that there has been a false detection. For
instance, assume that LDPC code blocks included in one PHY frame are assigned
block IDs representing sequential numbers in order. Here, if a block ID that is not
continuous with a block ID of the previous or next LDPC code block is found as a
result of the false detection check, then that block ID is judged as erroneous.
[0238]
The false detection check subunit 2101 outputs a result of the false detection
check in format that can be interpreted by software or hardware. For example, the
false detection check subunit 212 outputs "0" when no false detection has been
confirmed, and " 1" when a false detection has been confirmed. The result of the
false detection check output from the false detection check subunit 2101 is also a
result of the error detection output from the error detection unit 2003. On the other
hand, when the result of the parity check shows "1", it is obvious that the decoded
code word cd is not a code word of the LDPC code and is therefore erroneous. In this
case, there is no need to perform the false detection check. Therefore, in this case,
the false detection check subunit 2101 outputs "1" as a result of the false detection
check without performing the false detection check.
[0239]
As described above, in the present Embodiment 3, ID block bits that are
originally determined to be appended are used as false detection check bits. This
allows suppressing the throughput reduction caused by insertion of new false
detection check bits. Furthermore, in the present Embodiment 3, the false detection
check is performed by comparing the order based on numerical values obtained
through decoding of ID blocks with the order in which the ID blocks are actually
received. That is, one advantage of the present Embodiment 3 is that neither the
transmitter device nor the receiver device needs to determine what kind of bits are to
be used as false detection check bits (i.e., known bits).
[0240]
With the above-described structure, the decoding device 2000 can decode
the received code word and perform error detection on the result of the decoding.
The decoding device 2000 pertaining to the present Embodiment 3 is configured not
only to perform a parity check with use of the parity check matrix H of the LDPC
code, but also to check whether a code word obtained after the LDPC decoding is a
minimum-distance code word in relation to a corresponding error-free code word.
Accordingly, the decoding device 2000 pertaining to the present Embodiment 3 can
increase the error detection accuracy to a significant extent as compared to a
conventional error detection that only uses the parity check.
[0241]
As per the above Embodiment 2, when the number of block ID bits is
smaller than the number required as false detection check bits, extra false detection
check bits may be generated and inserted as described in the above Embodiment 1.
Supplementary Notes>
The above embodiments have described methods for implementing the
present invention. However, it goes without saying that the present invention is not
limited to these embodiments. The following describes various modification
examples that are included as the present invention in addition to the above
embodiments.
(1) The examples described in the above embodiments conform to the IEEE
802.1 In standard. However, communication conforming to a standard other than the
IEEE802.11n standard may be performed as long as the following aim is
accomplished: a false detection (i.e., judging an erroneous code word as an
error-free code word) is prevented by, in the course of QC-LDPC coding, inserting
bits for preventing the false detection (i.e., false detection check bits) into certain
positions in an information bit sequence to be transmitted, the certain positions
being different between (i) an original code word obtained by coding the
information bit sequence and (ii) a code word that is judged as an error-free code
word by the parity check but is different from the original code word with a close
Hamming distance therebetween. In this case, the parity check matrix to be used and
the false detection check bit insertion positions, which both vary depending on the
codeword length and code rate pertaining to the adopted standard, should be
calculated in advance.
(2) In the above embodiments, the insertion position determination table is
prestored in the receiver device. However, alternatively, the insertion position
determination table may be transmitted from the transmitter device to the receiver
device based on a conventional communication method. Note that in this case, the
insertion position determination table is transmitted prior to the data transfer
described in the above embodiments—i.e., transfer of data that has been coded with
false detection check bits inserted therein.
(3) As apparent from FIG. 4, a communication device pertaining to the
above embodiments is illustrated as if it has only one antenna for the sake of
convenience. However, a communication device of the present invention is not
limited to such structure; it may have multiple antennas.
(4) In the above Embodiments 1 through 3, the false detection check subunit
has been described as not performing the false detection check when the code word
subjected to LDPC decoding does not satisfy the parity check. This is simply
because there is no need to perform a false detection check. Alternatively, it is
permissible to perform a false detection check regardless of the result of the parity
check. In this case, a logical disjunction is performed on the result of the parity
check and the result of false detection check, and a value obtained from the logical
disjunction is output as the error detection result. FIG. 24 shows one exemplary
structure of error detection units 204 and 2003 that can perform the above logical
disjunction.
[0242]
A parity check subunit 241 and a false detection check subunit 242 shown
in FIG. 24 are substantially the same as the parity check subunit 211 and the false
detection check subunit 212 shown in FIG. 7, respectively. The parity check subunit
241 judges whether the decoded bit sequence that has been input satisfies the parity
check. The false detection check subunit 242 judges whether values held by the bits
located in the false detection check bit insertion positions are different from values
predetermined as the false detection check bits. The judgment results output from
the parity check subunit 241 and the false detection check subunit 242 are the same
as in the above Embodiment 1.
[0243]
Upon receiving values of the judgment results output from the parity check
subunit 241 and the false detection check subunit 242, a logical disjunction subunit
243 performs a logical disjunction on these values, and outputs a value obtained
from the logical disjunction as the error detection result.
[0244]
This concludes the description of one exemplary structure of the error
detection unit 204.
(5) In the above embodiments, a code word having the minimum Hamming
distance in relation to a transmitted code word is identified ahead of time, and
predetermined bit values for preventing a false detection are inserted into positions
that are different between the code word having the minimum Hamming distance
and the transmitted code word. Alternatively, such predetermined bit values for
preventing a false detection may be inserted into positions that are different between
a code word having the second or third closest Hamming distance in relation to a
transmitted code word, in addition to positions that are different between the code
word having the minimum Hamming distance and the transmitted code word. This
enables further improvement in the rate at which a false detection is prevented.
[0245]
In a case where an experiment or the like leads to the finding that a
transmitted code word easily changes into another code word that is different from a
code word having the minimum Hamming distance in relation to the transmitted
code word, the false detection check bits may be inserted into positions that have
different bit values between the transmitted code word and the stated code word that
easily changes into another code word, instead of positions that have different bit
values between the transmitted code word and the minimum-distance code word.
(6) With regard to each flowchart explained in the above embodiments
(with reference to, for example, FIGs. 9, 11, 16, 17, 22 and 23), the coding device
and decoding device do not necessarily have to operate in the order illustrated
therein, as long as the functional units of the coding device and decoding device
have received necessary information by the time when they execute their respective
processing. Therefore, parts of the processing performed by each of these functional
units may be executed in parallel when possible. It is also permissible to rearrange
the order of processing of each flowchart explained in the above embodiments. For
example, steps SI 104, SI 105 and SI 106 in FIG. 11 may be executed in parallel.
Alternatively, steps SI 105 and SI 106 may be executed before step SI 104. Similarly,
steps SI601 and SI602 in FIG. 16 may be executed in parallel. Alternatively, step
SI602 may be executed before step SI601.
(7) Each of the functional units included in the coding device and decoding
device illustrated in FIGs. 5, 6, 7, 14, 15, 19, 20 and 21 may be integrated and
realized as one or more LSIs (Large Scale Integrations). Alternatively, two or more
of such functional units may be collectively realized as a single LSI.
[0246]
Here, the LSI may be an IC (Integrated Circuit), a system LSI, a VLSI
(Very Large Scale Integration), an SLSI (Super Large Scale Integration), or a ULSI
(Ultra Large Scale Integration), depending on the degree of integration.
[0247]
Furthermore, the integration of circuits is not limited to being realized with
LSI, but may be realized with a special-purpose circuit or a general-use processor.
Alternatively, the integration may be realized with use of an FPGA (field
programmable gate array) that is programmable after manufacturing of the LSI, or a
re-configurable processor that enables re-configuration of the connection and
settings of circuit cells in the LSI.
[0248]
Furthermore, if technology for an integrated circuit that replaces LSIs
appears due to advances in or derivations from semiconductor technology, that
technology may be used for integration of functional blocks. Biotechnology is one
possible application.
(8) The operations and processing described in the above embodiments, such as
coding/decoding operations and processing of inserting false detection bits (for
example, see FIGs. 9, 11, 16, 17, 22 and 23), may be executed by processors
included in the transmitter device, receiver device, and the like, or by various
circuits connected to such processors. A control program composed of program
codes for causing the processors and various circuits to execute such operations and
processing may be recorded on a recording medium, or distributed/disseminated via
different types of communication channels. Examples of such a recording medium
include an IC card, a hard disk, an optical disc, a flexible disk, and a ROM. Once the
distributed/disseminated control program is stored into memory or the like that can
be read by processors, the control program can be provided for use. Here, the
various functions described in the above embodiments can be realized by the
processors executing the control program.
[Industrial Applicability]
[0249]
A coding device and a decoding device pertaining to the present invention
can reduce the rate of false detection whereby an erroneous code word is judged as
an error-free code word during a QC-LDPC parity check, without reducing the
throughput of data transfer as compared to a case where a CRC code or the like is
used. Accordingly, an effective use of such coding device and decoding device can
be expected in communication systems.
[Reference Signs List]
[0250]
100, 1400, 1900 coding device
101 LDPC coding parameter storage unit
102, 1401, 1901 false detection check bit control unit
103 false detection check bit insertion unit
104 LDPC coding unit
200, 1500, 2000 decoding device
201 LDPC coding parameter storage unit
202 LDPC decode unit
203 false detection check bit control unit
204, 2003 error detection unit
205 false detection check bit removal unit
211 parity check subunit
212,2101 false detection check subunit
1402 shortening bit insertion unit
1403 rearrangement unit
1404 shortening bit removal unit
1501 shortening bit insertion unit
1502 rearrangement unit
1503 rearrangement unit
15 04 shortening bit removal unit
2001 parity bit removal unit
2002 rearrangement unit

WE CLAIM:

1. A coding device for coding an information bit sequence and outputting the coded
information bit sequence, comprising:
a determining unit operable to, in accordance with an applied coding
method that utilizes a quasi-cyclic code, determine at least one bit position in the
information bit sequence as an insertion position into which a predetermined bit
value is to be inserted, wherein a bit position in a first code word which corresponds
to the insertion position holds a different bit value from a bit position in at least one
second code word which corresponds to the insertion position, the first code word
being obtained by coding the information bit sequence in accordance with the
coding method, the second code word being judged as an error-free code word by a
parity check of the coding method but being different from the first code word by
the minimum Hamming distance;
an inserting unit operable to generate a pre-coding bit sequence by inserting
the predetermined bit value into the insertion position of the information bit
sequence; and
a coding unit operable to generate a third code word by coding the
pre-coding bit sequence in accordance with the coding method.
2. The coding device of Claim 1, further comprising
a known bit appending unit operable to append at least one known bit,
which has been predetermined as a shortening bit of a shortening code, to the
information bit sequence when an information length of the information bit
sequence is shorter than an information length of an information bit portion of a
code word determined based on a codeword length and a code rate pertaining to the
coding method, wherein
the inserting unit rearranges the information bit sequence to which the
known bit has been appended, so that the known bit is inserted into the insertion
position as the predetermined bit value.
3. The coding device of Claim 2, further comprising
a removing unit operable to remove one or more of the at least one known
bit appended by the known bit appending unit from the third code word.
4. The coding device of Claim 1, wherein
the third code word generated by the coding unit includes a block ID for
identifying the third code word among other third code words, and
the inserting unit inserts a bit value of a bit representing the block ID into
the insertion position as the predetermined bit value.
5. A decoding device comprising:
a decoding unit operable to generate a decoded bit sequence by performing
decode processing on a pre-decoding bit sequence input thereto, the decode
processing complying with an applied coding method that utilizes a quasi-cyclic
code;
a checking unit operable to check whether the decoded bit sequence is
judged as an error-free code word by a parity check of the coding method;
a judging unit operable to judge whether at least one insertion position in
the decoded bit sequence holds at least one predetermined bit value, wherein
according to the coding method, a bit position in a first code word which
corresponds to the insertion position is predetermined to hold a different bit value
from a bit position in a second code word which corresponds to the insertion
position, the first code word being generated by a transmitter device according to the
coding method, and the second code word being judged as an error-free code word
by the parity check but being different from the first code word by the minimum
Hamming distance; and
an error detecting unit operable to, in a case where the judging unit judges
in the negative, judge that the decoded bit sequent is erroneous even when a result of
the checking by the checking unit shows that the decoded bit sequence has been
judged as the error-free code word by the parity check.
6. The decoding device of Claim 5, further comprising
an inserting unit operable to, when a pre-decoding shortened code word
from which one or more of the at least one predetermined bit value have been
removed by the transmitter device is input thereto, generate the pre-decoding bit
sequence by inserting at least one known bit holding a known value into the input
pre-decoding shortened code word.
7. The decoding device of Claim 5, wherein
the judging unit judges whether the insertion position in the decoded bit
sequence holds a bit value that can be predicted from a predetermined bit value of
another decoded bit sequence, instead of judging whether the insertion position in
the decoded bit sequence holds the at least one predetermined bit value.
8. A coding method for coding an information bit sequence, comprising the steps of:
(A) determining, in accordance with an applied coding method that utilizes
a quasi-cyclic code, at least one bit position in the information bit sequence as an
insertion position into which a predetermined bit value is to be inserted, wherein a
bit position in a first code word which corresponds to the insertion position holds a
different bit value from a bit position in at least one second code word which
corresponds to the insertion position, the first code word being obtained by coding
the information bit sequence in accordance with the coding method, the second code
word being judged as an error-free code word by a parity check of the coding
method but being different from the first code word by the minimum Hamming
distance;
(B) generating a pre-coding bit sequence by inserting the predetermined bit
value into the insertion position of the information bit sequence; and
(C) generating a third code word by coding the pre-coding bit sequence in
accordance with the coding method.
9. A decoding method comprising the steps of:
(A) generating a code word by decoding a received signal in accordance
with an applied coding method that utilizes a quasi-cyclic code;
(B) checking whether the code word generated in the step (A) is an
error-free code word by a parity check of the coding method;
(C) judging whether at least one insertion position in the generated code
word holds at least one predetermined bit value, wherein according to the coding
method, a bit position in a first code word which corresponds to the insertion
position is predetermined to hold a different bit value from a bit position in a second
code word which corresponds to the insertion position, the first code word being
generated by a transmitter device according to the coding method, and the second
code word being judged as an error-free code word by the parity check but being
different from the first code word by the minimum Hamming distance; and
(D) in a case where the judgment in the step (C) is in the negative, judging
that the generated code word is erroneous even when a result of the checking by the
step (B) shows that the generated code word has been judged as the error-free code
word by the parity check.
10. A communication system including a transmitter device and a receiver device,
the transmitter device, which codes an information bit sequence and
transmits the coded information bit sequence, comprising:
a determining unit operable to, in accordance with an applied coding
method that utilizes a quasi-cyclic code, determine at least one bit position in the
information bit sequence as an insertion position into which a predetermined bit
value is to be inserted, wherein a bit position in a first code word which corresponds
to the insertion position holds a different bit value from a bit position in at least one
second code word which corresponds to the insertion position, the first code word
being obtained by coding the information bit sequence in accordance with the
coding method, the second code word being judged as an error-free code word by a
parity check of the coding method but being different from the first code word by
the minimum Hamming distance;
an inserting unit operable to generate a pre-coding bit sequence by inserting
the predetermined bit value into the insertion position of the information bit
sequence;
a coding unit operable to generate a third code word by coding the
pre-coding bit sequence in accordance with the coding method; and
a transmitting unit operable to transmit a wireless signal containing the third
code word,
the receiver device comprising:
a receiving unit operable to receive the wireless signal;
a decoding unit operable to generate a code word by decoding the received
signal in accordance with the coding method;
a checking unit operable to check whether the code word generated by the
decoding unit is judged as an error-free code word by the parity check of the coding
method;
a judging unit operable to judge whether at least one insertion position in
the generated code word holds at least one predetermined bit value, wherein
according to the coding method, a bit position in a first code word which
corresponds to the insertion position is predetermined to hold a different bit value
from a bit position in a second code word which corresponds to the insertion
position, the first code word being generated by the transmitter device according to
the coding method, and the second code word being judged as an error-free code
word by the parity check but being different from the first code word by the
minimum Hamming distance; and
an error detecting unit operable to, in a case where the judging unit judges
in the negative, judge that the generated code word is erroneous even when a result
of the checking by the checking unit shows that the generated code word has been
judged as the error-free code word by the parity check.

Positions holding different bit values between a first code word, which is
obtained by coding an information bit sequence based on a coding method utilizing
quasi-cyclic codes, and a second code word, which has the close Hamming distance
from the first code word and satisfies a parity check of the coding method, are
identified. Thereafter, a code word is generated by inserting bit values known to the
transmitter and receiver into the identified positions of the information bit sequence
and coding the information bit sequence. Upon reception of a signal based on the
generated code word, the receiver judges whether known bit values held by
corresponding positions in a code word obtained by decoding the received signal are
the same as preset bit values. If the judgment result is negative, the code word based
on the received signal is judges as erroneous even when it satisfies the parity check.