FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(See section 10, rule 13)
“RECEPTION DEVICE, COMMUNICATION SYSTEM,
AND METHOD FOR CALCULATING LIKELIHOOD OF
MODULATION SIGNAL”
MITSUBISHI ELECTRIC CORPORATION, of 7-3,
Marunouchi 2-chome, Chiyoda-ku, Tokyo 100-8310 Japan
The following specification particularly describes the invention and the manner in
which it is to be performed.
2
Field
[0001] The present invention relates to a reception
device that receives a multiplexed signal, a communication
system including the reception device, and a method for
5 calculating a likelihood of a modulation signal which is
applied to the reception device.
Background
[0002] In communication to which a multiband OFDM
10 (Orthogonal Frequency Division Multiplexing) method is
applied, a plurality of modulation signals are multiplexed
by a precoding matrix or the like and then transmitted. In
communication in which signals are multiplexed on the
transmitter side, such as communication using the multiband
15 OFDM method, the multiplexed signals need to be separated
from each other on the receiver side. As a signal
separation method, an MLD (Maximum Likelihood Detection)
method is exemplified. In the MLD method, signal
separation is performed by obtaining a distance between a
20 reception signal vector and each of candidate signal points
and determining a signal point at the shortest distance
from the reception signal vector as an estimated signal
vector. Non Patent Literature 1 discloses a method to
reduce the amount of computation in the MLD method. In the
25 disclosed method, a real component and an imaginary
component of a signal are independently determined, and a
signal with a real component or imaginary component assumed
at a candidate signal point is used to sequentially
estimate a real component or an imaginary component of the
30 remaining signals. In that case, the disclosed method
reduces the number of candidate signal points to be used
for the assumption on the basis of a result of region
determination using a reception signal.
3
Citation List
Non Patent Literature
[0003] Non Patent Literature 1: YAMAGUCHI KANAKO,
5 NISHIMOTO HIROSHI, UMEDA SHUSAKU, TSUKAMOTO KAORU, OKAZAKI
AKIHIRO, SANO HIROYASU, OKAMURA ATSUSHI, "A Study on
Reduction in Candidate Signal Points of MLD Decoding in
Frequency Encoded Diversity Method", 2016 IEICE Society
Conference, B-5-20, p.290, 2016.
10
Summary
Technical Problem
[0004] However, the method to reduce the amount of
computation in the MLD method described in Non Patent
15 Literature 1 has a problem that as the number of
multiplexed signals increases, the amount of computation
increases.
[0005] The present invention has been achieved in view
of the above problems, and an object of the present
20 invention is to provide a reception device that can reduce
the amount of computation in a signal separation process
even when the number of multiplexed signals increases.
Solution to Problem
25 [0006] In order to solve the above problems and achieve
the object, a reception device according to the present
invention comprises: a signal division unit to divide a
reception signal including a plurality of multiplexed
signals respectively into a real component and an imaginary
30 component, the multiplexed signals being obtained by
multiplexing a plurality of modulation signals by a realnumber
precoding matrix, each of the modulation signals
having a real component and an imaginary component
4
modulated independently from each other; a first maximumlikelihood
point search unit to narrow down candidate
signal points, which are obtainable by a real component of
the multiplexed signal, to a first candidate signal point
5 by using one of the real components of the reception
signal; a second maximum-likelihood point search unit to
narrow down candidate signal points, which are obtainable
by an imaginary component of the multiplexed signal, to a
second candidate signal point by using one of the imaginary
10 components of the reception signal; a first replica-vector
calculation unit to calculate a first replica vector by
using the first candidate signal point; a second replicavector
calculation unit to calculate a second replica
vector by using the second candidate signal point; and a
15 likelihood calculation unit to calculate a likelihood of
the modulation signal by using the first replica vector and
the second replica vector.
Advantageous Effects of Invention
20 [0007] The reception device according to the present
invention has an effect where it is possible to reduce the
amount of computation in a signal separation process even
when the number of multiplexed signals increases.
25 Brief Description of Drawings
[0008] FIG. 1 is a diagram illustrating a communication
system according to an embodiment.
FIG. 2 is a functional block diagram of a signal
detection unit according to the embodiment.
30 FIG. 3 is a diagram illustrating an example of a real
component Re(y1) of a complex baseband signal input to a
maximum-likelihood point search unit according to the
embodiment.
5
FIG. 4 is a diagram illustrating a configuration
example of a control circuit according to the embodiment.
FIG. 5 is a flowchart illustrating an example of
processes in a reception device according to the
5 embodiment.
Description of Embodiment
[0009] A reception device, a communication system, and a
method for calculating a likelihood of a modulation signal
10 according to an embodiment of the present invention will be
described in detail below with reference to the
accompanying drawings. The present invention is not
limited to the embodiment.
[0010] Embodiment.
15 FIG. 1 is a diagram illustrating a communication
system according to an embodiment. A communication system
1 includes a transmission device 10 and a reception device
20. The transmission device 10 includes a precoding unit
100. The reception device 20 includes a signal detection
20 unit 200. The precoding unit 100 generates a transmission
signal by performing a modulation process and a precoding
process on information signals s1, s2, and s3 to be
transmitted to the reception device 20. The precoding unit
100 transmits the transmission signal to the reception
25 device 20 through propagation paths 30a, 30b, and 30c. The
reception device 20 decodes the information signals s1, s2,
and s3 with the signal detection unit 200 performing a
signal separation process on the reception signal. In the
present embodiment, the transmission device 10 multiplexes
30 a plurality of modulated complex baseband signals by using
a real-number precoding matrix, and then transmits the
multiplexed signals through propagation paths orthogonal to
each other. The propagation paths orthogonal to each other
6
refer to propagation paths that are less likely to
interfere with each other, or refer to independent
propagation paths that do not interfere with each other.
While examples of the propagation paths orthogonal to each
5 other include propagation paths using orthogonal
frequencies, the propagation paths orthogonal to each other
are not limited thereto.
[0011] An operation of the transmission device 10 is
described below in detail. The precoding unit 100 performs
10 a modulation process and a multiplexing process on the
information signals s1, s2, and s3 received respectively
through signal lines s100a, s100b, and s100c. It is
assumed that there are three signals to be multiplexed in
the multiplexing process performed by the precoding unit
15 100 according to the present embodiment. Each of the three
information signals, input to the precoding unit 100, is
modulated by a QPSK (Quadrature Phase Shift Keying) method.
The present embodiment is not limited to the QPSK method,
but is also applicable to a case where a real component and
20 an imaginary component of a complex baseband signal are
modulated independently from each other. In other words,
the present embodiment is applicable to a modulation method
that can calculate the real component and the imaginary
component independently from each other. The number M of
25 signals to be multiplexed in the multiplexing process
performed by the precoding unit 100 is not limited to three
and it suffices that the number is an integer equal to or
larger than 2. The information signals s1, s2, and s3 are,
for example, information such as (01), (00), or (11). The
30 precoding unit 100 performs a modulation process on each of
the information signals s1, s2, and s3 to generate
respective complex baseband signals that are modulation
signals z1, z2, and z3. The information signals s1, s2,
7
and s3 uniquely correspond to the modulation signals z1,
z2, and z3. That is, the information signal s1 corresponds
to the modulation signal z1, the information signal s2
corresponds to the modulation signal z2, and the
5 information signal s3 corresponds to the modulation signal
z3. In the following descriptions, either the information
signals s1, s2, and s3 or the modulation signals z1, z2,
and z3 are used for explanation.
[0012] The precoding unit 100 performs a multiplexing
10 process on the modulation signals z1, z2, and z3 on the
basis of a real-number precoding matrix included in the
precoding unit 100. Three multiplexed radio signals are
output from the precoding unit 100 to the propagation paths
30a, 30b, and 30c that are orthogonal to each other. In
15 the precoding unit 100, the modulation signals z1, z2, and
z3 are multiplexed by a real-number precoding matrix in
which the amount of phase rotation becomes an integral
multiple of 90 degrees. In a case where the modulation
signals z1, z2, and z3 are transmitted through three
20 propagation paths, the real-number precoding matrix refers
to a matrix having already defined therein the mixture
ratio of the modulation signals z1, z2, and z3 on their
respective propagation paths. The real-number precoding
matrix is shared by the transmission device 10 and the
25 reception device 20. The reception device 20 can use the
real-number precoding matrix when the reception device 20
decodes a reception signal.
[0013] The precoding unit 100 performs a multiplexing
process on a complex baseband signal to be transmitted by
30 multiplying a modulation signal vector z, which is a vector
value of the complex baseband signal, by a real-number
precoding matrix , and then transmits the multiplexed
signal to a propagation path. That is, a transmission
8
signal vector x that is output by the precoding unit 100 is
expressed by the following equation.
[0014]
[Equation 1]
5 •••(1)
[0015] When the transmission signal vector x passes
through the propagation paths 30a, 30b, and 30c, the
transmission signal vector x is influenced by each of the
propagation paths 30a, 30b, and 30c. The influence upon
10 the transmission signal vector x can be expressed by a
transfer function matrix Δ. The transfer function matrix Δ
can be estimated by the transmission device 10, the
reception device 20, or other devices (not illustrated).
The reception device 20 has information about this transfer
15 function matrix Δ. A noise vector η at an input terminal
of the reception device 20 is further added to the
transmission signal vector x.
[0016] A reception signal vector y, which has been input
to the reception device 20 after having been influenced by
20 the propagation paths and noise, is made up of complex
baseband signals that are complex baseband signals y1, y2,
and y3. The reception signal vector y can be expressed by
a complex vector with the number of dimensions equal to the
number of propagation paths orthogonal to each other. The
25 complex baseband signal y1 is input to the reception device
20 through the propagation path 30a. The complex baseband
signal y2 is input to the reception device 20 through the
propagation path 30b. The complex baseband signal y3 is
input to the reception device 20 through the propagation
30 path 30c. In the present embodiment, the number of
propagation paths orthogonal to each other is three. The
x = ��z
��
��1
��2
��3
�� = ��
��11 ��12 ��13
��21 ��22 ��23
��31 ��32 ��33
�� ��
��1
��2
��3
��
9
reception signal vector y can be expressed by the following
equation using: the real-number precoding matrix by
which the modulation signal vector z is multiplied in the
transmission device 10, the transfer function matrix Δ of
5 the propagation paths estimated in the reception device 20
and the transmission device 10, the modulation signal
vector z, and the noise vector η added at the input
terminal of the reception device 20.
[0017]
10 [Equation 2]
•••(2)
[0018] An operation of the reception device 20 is
described below in detail. The reception signal vector y
is input to the signal detection unit 200. The reception
15 device 20 performs a process to derive the transmitted
modulation signal vector z from the reception signal vector
y. The signal detection unit 200 has a function of
performing signal separation on three input radio signals.
The signal detection unit 200 estimates the three separated
20 radio signals, and outputs a likelihood of each of the
estimated signals.
[0019] FIG. 2 is a functional block diagram of the
signal detection unit 200 according to the embodiment. The
signal detection unit 200 includes: a signal division unit
25 210; first maximum-likelihood point search units 220a to
220c; second maximum-likelihood point search units 221a to
221c; first replica-vector calculation units 230a to 230c;
second replica-vector calculation units 231a to 231c; and a
likelihood calculation unit 240.
30 [0020] The signal division unit 210 divides each of the
�� = ���� + ��
��
��1
��2
��3
�� = ��
Δ1 0 0
0 Δ2 0
0 0 Δ3
�� ��
��11 ��12 ��13
��21 ��22 ��23
��31 ��32 ��33
�� ��
��1
��2
��3
�� + ��
10
complex baseband signals y1, y2, and y3 of the reception
signal vector y into a real component and an imaginary
component. The reception signal vector y is input to the
signal division unit 210 through the signal lines s200a,
5 s200b, and s200c. The complex baseband signal y1 is input
to the signal division unit 210 through the signal line
s200a. The complex baseband signal y2 is input to the
signal division unit 210 through the signal line s200b.
The complex baseband signal y3 is input to the signal
10 division unit 210 through the signal line s200c. The
signal division unit 210 outputs a real component of the
complex baseband signal y1 to the first maximum-likelihood
point search unit 220a, and outputs an imaginary component
of the complex baseband signal y1 to the second maximum15
likelihood point search unit 221a. The signal division
unit 210 outputs a real component of the complex baseband
signal y2 to the first maximum-likelihood point search unit
220b, and outputs an imaginary component of the complex
baseband signal y2 to the second maximum-likelihood point
20 search unit 221b. The signal division unit 210 outputs a
real component of the complex baseband signal y3 to the
first maximum-likelihood point search unit 220c, and
outputs an imaginary component of the complex baseband
signal y3 to the second maximum-likelihood point search
25 unit 221c.
[0021] The first maximum-likelihood point search unit
220a uses the real component of the complex baseband signal
y1 to narrow down candidate signal points, which are
obtainable by a real component of a multiplexed signal x1
30 that is multiplexed by a real-number precoding matrix, to a
candidate signal point that is located at the shortest
distance from the real component of the complex baseband
signal y1. The distance used to narrow down the candidate
11
signal points is the Euclidean distance. The second
maximum-likelihood point search unit 221a uses the
imaginary component of the complex baseband signal y1 to
narrow down candidate signal points, which are obtainable
5 by an imaginary component of the multiplexed signal x1 that
is multiplexed by a real-number precoding matrix, to a
candidate signal point that is located at the shortest
distance from the imaginary component of the complex
baseband signal y1. Similarly, the first maximum10
likelihood point search unit 220b uses the real component
of the complex baseband signal y2 to narrow down candidate
signal points, which are obtainable by a real component of
a multiplexed signal x2, to a candidate signal point that
is located at the shortest distance from the real component
15 of the complex baseband signal y2. The second maximumlikelihood
point search unit 221b uses the imaginary
component of the complex baseband signal y2 to narrow down
candidate signal points, which are obtainable by an
imaginary component of the multiplexed signal x2, to a
20 candidate signal point that is located at the shortest
distance from the imaginary component of the complex
baseband signal y2. The first maximum-likelihood point
search unit 220c uses the real component of the complex
baseband signal y3 to narrow down candidate signal points,
25 which are obtainable by a real component of a multiplexed
signal x3, to a candidate signal point that is located at
the shortest distance from the real component of the
complex baseband signal y3. The second maximum-likelihood
point search unit 221c uses the imaginary component of the
30 complex baseband signal y3 to narrow down candidate signal
points, which are obtainable by an imaginary component of
the multiplexed signal x3, to a candidate signal point that
is located at the shortest distance from the imaginary
12
component of the complex baseband signal y3.
[0022] The first maximum-likelihood point search units
220a to 220c output the candidate signal point having been
narrowed down to the first replica-vector calculation units
5 230a to 230c, respectively. The second maximum-likelihood
point search units 221a to 221c output the candidate signal
point having been narrowed down to the second replicavector
calculation units 231a to 231c, respectively. For
example, the first maximum-likelihood point search unit
10 220a outputs the candidate signal point to the first
replica-vector calculation unit 230a. For example, the
second maximum-likelihood point search unit 221c outputs
the candidate signal point to the second replica-vector
calculation unit 231c. The candidate signal point narrowed
15 down by the first maximum-likelihood point search units
220a to 220c is also referred to as "first candidate signal
point". The candidate signal point narrowed down by the
second maximum-likelihood point search units 221a to 221c
is also referred to as "second candidate signal point".
20 [0023] The first replica-vector calculation units 230a
to 230c and the second replica-vector calculation units
231a to 231c calculate a plurality of replica vectors
corresponding to the modulation signal z1, the modulation
signal z2, or the modulation signal z3, which are
25 calculated using an input maximum likelihood point. The
replica vectors calculated by the first replica-vector
calculation units 230a to 230c are also referred to as
"first replica vector". The replica vectors calculated by
the second replica-vector calculation units 231a to 231c
30 are also referred to as "second replica vector". The first
replica-vector calculation units 230a to 230c output the
replica vectors calculated using the maximum likelihood
point, and a plurality of vectors to the likelihood
13
calculation unit 240 as a group of replica vectors. The
plurality of vectors are made up of a candidate signal
point, located at the shortest distance from a real
component of the complex baseband signal y1, of the complex
5 baseband signal y2, or of the complex baseband signal y3
having been respectively input to the first maximumlikelihood
point search units 220a to 220c, among candidate
signal points with an inverted value at each bit of the
maximum likelihood point. Similarly, the second replica10
vector calculation units 231a to 231c output the replica
vectors calculated using the maximum likelihood point, and
a plurality of vectors to the likelihood calculation unit
240 as a group of replica vectors. The plurality of
vectors are made up of a candidate signal point, located at
15 the shortest distance from an imaginary component of the
complex baseband signal y1, of the complex baseband signal
y2, or of the complex baseband signal y3 having been
respectively input to the second maximum-likelihood point
search units 221a to 221c, among candidate signal points
20 with an inverted value at each bit of the maximum
likelihood point.
[0024] For example, the first replica-vector calculation
unit 230a uses a maximum likelihood point located at the
shortest distance from the real component of the complex
25 baseband signal y1 to calculate a replica vector
corresponding to the multiplexed signal x1. The first
replica-vector calculation unit 230a outputs the calculated
replica vector and a plurality of vectors to the likelihood
calculation unit 240 as a group of replica vectors. The
30 plurality of vectors are made up of a candidate signal
point, which is located at the shortest distance from a
real component of the complex baseband signal y1 having
been input to the first maximum-likelihood point search
14
unit 220a, among candidate signal points with an inverted
value at each bit of the maximum likelihood point.
[0025] The likelihood calculation unit 240 calculates a
likelihood corresponding to each of the modulation signals
5 z1, z2, and z3 using a plurality of input replica vectors.
The likelihood calculation unit 240: outputs a likelihood
corresponding to the modulation signal z1 through the
signal line s201a; outputs a likelihood corresponding to
the modulation signal z2 through the signal line s201b; and
10 outputs a likelihood corresponding to the modulation signal
z3 through the signal line s201c.
[0026] An operation of the signal detection unit 200 is
described below in detail. Initially, the first maximumlikelihood
point search units 220a to 220c and the second
15 maximum-likelihood point search units 221a to 221c narrow
down candidate signal points, which are obtainable by a
real component and an imaginary component of each of the
multiplexed signals x1, x2, and x3, to a candidate signal
point located at the shortest distance from each of the
20 complex baseband signals y1, y2, and y3 using a real
component and an imaginary component of each of the input
complex baseband signals y1, y2, and y3.
[0027] The present embodiment deals with a case where
the QPSK method is applied as a modulation method for the
25 modulation signals z1, z2, and z3. Thus, all the candidate
signal points that are obtainable by real components
Re(z1), Re(z2), and Re(z3) of the modulation signals are
2×2×2=8 types ([z1, z2, z3]=[0, 0, 0], [0, 0, 1], [0, 1,
1], [0, 1, 0], [1, 0, 0], [1, 0, 1], [1, 1, 1], [1, 1, 0]).
30 FIG. 3 illustrates locations of signal points where the
horizontal axis represents a real component Re(y1) of a
complex baseband signal.
[0028] FIG. 3 is a diagram illustrating an example of
15
the real component Re(y1) of a complex baseband signal
input to the first maximum-likelihood point search unit
220a according to the embodiment. In FIG. 3, the black
spot illustrates a real component Re(x1) of a multiplexed
5 signal. When the real component Re(y1) is illustrated as a
point marked with "×", a candidate point [0, 0, 0] for the
real component Re(x1) of the multiplexed signal is located
at the shortest distance from the real component Re(y1).
For this reason, the candidate point [0, 0, 0] is optimal
10 as a maximum likelihood point of the real component Re(y1).
Thus, the first maximum-likelihood point search unit 220a
outputs the signal point [0, 0, 0] as a maximum likelihood
point to the first replica-vector calculation unit 230a.
[0029] In addition to the maximum likelihood point,
15 information about an inverted bit at each bit of the
maximum likelihood point is necessary for the likelihood
calculation unit 240 to output a likelihood of a modulation
signal. In addition to the maximum likelihood point [0, 0,
0] input from the first maximum-likelihood point search
20 unit 220a, the first replica-vector calculation unit 230a
also outputs candidate signal points [0, 1, 0], [1, 0, 0],
and [0, 0, 1], which are located at the shortest distance
from the maximum likelihood point, among candidate signal
points including an inverted bit at each bit of the maximum
25 likelihood point, as replica vectors to the likelihood
calculation unit 240.
[0030] In the present embodiment, the candidate signal
points are narrowed down to determine which one of them
becomes the maximum likelihood point for the real component
30 Re(y1) of a complex baseband signal by determining in which
of the regions, defined by dotted lines illustrated on the
signal-point location diagram in FIG. 3, the real component
Re(y1) of the complex baseband signal is included. The
16
regions can be calculated on the basis of locations of the
candidate signal points. For example, it is possible to
calculate the regions defined by the dotted lines on the
basis of the distance between the adjacent candidate signal
5 points of the real component Re(x1) of the multiplexed
signal x1 that is multiplexed by a real-number precoding
matrix. It is also possible to calculate which of the
candidate signal points is selected as an inverted bit to
the maximum likelihood point for the real component Re(y1)
10 of the complex baseband signal, on the basis of locations
of the signal points of the real component Re(x1) of the
multiplexed signal illustrated in FIG. 3. It is allowable
that the first maximum-likelihood point search units 220a
to 220c, the second maximum-likelihood point search units
15 221a to 221c, the first replica-vector calculation units
230a to 230c, and the second replica-vector calculation
units 231a to 231c do not calculate information about the
regions defined by the dotted lines or information about
the locations of candidate signal points. For example, the
20 transmission device 10 may calculate information about the
regions defined by the dotted lines, and information about
the locations of candidate signal points, so that the
calculated information may be shared between the
transmission device 10 and the reception device 20.
25 [0031] Similarly, the first maximum-likelihood point
search units 220b and 220c, and the second maximumlikelihood
point search units 221a to 221c search a maximum
likelihood point on the basis of the real components Re(y2)
and Re(y3) and imaginary components Im(y1), Im(y2), and
30 Im(y3) of respective complex baseband signals. The first
replica-vector calculation units 230b and 230c, and the
second replica-vector calculation units 231a to 231c
calculate a replica vector on the basis of the searched
17
maximum likelihood point.
[0032] The present embodiment deals with a case where
three modulation signals modulated by the QPSK method are
multiplexed by a real-number precoding matrix. Thus, each
5 of the six replica-vector calculation units outputs four
replica vectors. Accordingly, 24 replica vectors are
output in total. Similarly, in a case where M modulation
signals, obtained by modulating an information signal of N
bits using a modulation method for modulating a real
10 component and an imaginary component independently, are
multiplexed by a real-number precoding matrix, then each of
2M replica-vector calculation units outputs (NM/2+1)
replica vectors. Accordingly, (NM2+2M) replica vectors are
output in total. In a case where a modulation signal has a
15 value of only the real component or a value of only the
imaginary component as modulated by the BPSK (Binary Phase
Shift Keying) method, each of M replica-vector calculation
units outputs (NM/2+1) replica vectors. Accordingly,
(NM2/2+M) replica vectors are output in total.
20 [0033] The likelihood calculation unit 240 calculates a
likelihood corresponding to each bit of the modulation
signals z1, z2, and z3 using all the replica vectors
calculated by the first replica-vector calculation units
230a to 230c and the second replica-vector calculation
25 units 231a to 231c. The likelihood calculation unit 240
outputs the calculated likelihoods respectively through the
signal lines s201a, s201b, and s201c. Likelihood
calculation can use the existing method in which the
probability of occurrence of 0 and 1 at each bit is
30 calculated on the basis of the shortest distance from a
reception signal vector.
[0034] Descriptions are made of a hardware configuration
of the signal detection unit 200, the signal division unit
18
210, the first maximum-likelihood point search units 220a
to 220c, the second maximum-likelihood point search units
221a to 221c, the first replica-vector calculation units
230a to 230c, the second replica-vector calculation units
5 231a to 231c, and the likelihood calculation unit 240
according to the present embodiment. FIG. 4 is a diagram
illustrating a configuration example of a control circuit
according to the embodiment. The signal detection unit
200, the signal division unit 210, the first maximum10
likelihood point search units 220a to 220c, the second
maximum-likelihood point search units 221a to 221c, the
first replica-vector calculation units 230a to 230c, the
second replica-vector calculation units 231a to 231c, and
the likelihood calculation unit 240 are implemented by a
15 processing circuit that is an electronic circuit to perform
each process.
[0035] It is allowable that this processing circuit is
either dedicated hardware, or a control circuit including a
memory and a CPU (Central Processing Unit) that executes a
20 program stored in the memory. For example, the memory
described herein is a nonvolatile or volatile semiconductor
memory such as a RAM (Random Access Memory), a ROM (Read
Only Memory), or a flash memory, or is a magnetic disk or
an optical disk. In a case where this processing circuit
25 is a control circuit including the CPU, this control
circuit is, for example, a control circuit 300 that is
configured as illustrated in FIG. 4.
[0036] As illustrated in FIG. 5, the control circuit 300
includes a processor 300a that is a CPU, and a memory 300b.
30 In a case where the processing circuit is implemented by
the control circuit 300 illustrated in FIG. 5, the
processor 300a reads and executes a program that is stored
in the memory 300b, and that corresponds to each process,
19
thereby implementing the processing circuit. The memory
300b is also used as a temporary memory for the processor
300a to perform each process.
[0037] A process flow of the reception device 20
5 according to the present embodiment is described below.
FIG. 5 is a flowchart illustrating an example of the
processes in the reception device 20 according to the
present embodiment.
[0038] The signal detection unit 200 receives the
10 reception signal vector y through the signal lines s200a,
s200b, and s200c (Step S101).
[0039] When the signal detection unit 200 receives the
reception signal vector y, the signal division unit 210
divides each of the complex baseband signals y1, y2, and y3
15 making up the reception signal vector y into a real
component and an imaginary component. The signal division
unit 210: outputs a real component of the complex baseband
signal y1 to the first maximum-likelihood point search unit
220a; outputs an imaginary component of the complex
20 baseband signal y1 to the second maximum-likelihood point
search unit 221a; outputs a real component of the complex
baseband signal y2 to the first maximum-likelihood point
search unit 220b; outputs an imaginary component of the
complex baseband signal y2 to the second maximum-likelihood
25 point search unit 221b; outputs a real component of the
complex baseband signal y3 to the first maximum-likelihood
point search unit 220c; and outputs an imaginary component
of the complex baseband signal y3 to the second maximumlikelihood
point search unit 221c (Step S102). Step S102
30 is also referred to as "first step".
[0040] The first maximum-likelihood point search units
220a to 220c and the second maximum-likelihood point search
units 221a to 221c narrow down candidate signal points,
20
which are obtainable by a signal that is one of the
components of the transmission signal vector x multiplexed
by a real-number precoding matrix, to a candidate signal
point located at the shortest distance from the reception
5 signal vector y, and then output the candidate signal point
to the first replica-vector calculation units 230a to 230c
and the second replica-vector calculation units 231a to
231c (Step S103). Step S103 performed by the first
maximum-likelihood point search units 220a to 220c is also
10 referred to as "second step". Step S103 performed by the
second maximum-likelihood point search units 221a to 221c
is also referred to as "third step".
[0041] On the basis of the candidate signal point, the
first replica-vector calculation units 230a to 230c and the
15 second replica-vector calculation units 231a to 231c
calculate a plurality of replica vectors having a maximum
likelihood point or having an inverted bit to the candidate
signal point, and then output the calculated replica
vectors to the likelihood calculation unit 240 (Step S104).
20 Step S104 performed by the first replica-vector calculation
units 230a to 230c is also referred to as "fourth step".
Step S104 performed by the second replica-vector
calculation units 231a to 231c is also referred to as
"fifth step".
25 [0042] The likelihood calculation unit 240 calculates a
likelihood of the modulation signal vector z using a
plurality of replica vectors output from the first replicavector
calculation units 230a to 230c and the second
replica-vector calculation units 231a to 231c (Step S105).
30 Step S105 is also referred to as "sixth step".
[0043] As described above, in the present embodiment,
the reception device 20 performs signal separation on a
real component and an imaginary component of a reception
21
signal. Particularly in the signal separation, on the
basis of a value of each component of the reception signal,
candidate signal points, which are obtainable by a
multiplexed signal that is multiplexed by a real-number
5 precoding matrix, are narrowed down independently to a
candidate signal point located at the shortest distance
from the reception signal vector. Then, a replica vector
is calculated from this candidate signal point having been
narrowed down, and the calculated replica vector is used to
10 calculate a likelihood. This makes it possible for the
reception device 20 to decode the frequency with a smaller
amount of computation even when a greater number of signals
are multiplexed.
[0044] It is also allowable that the signal detection
15 unit 200 is provided with maximum-likelihood point search
units and replica-vector calculation units corresponding to
the possible maximum number of signals to be multiplexed,
so that the signal detection unit 200 adjusts the number of
maximum-likelihood point search units and replica-vector
20 calculation units to be used in accordance with the number
of signals to be multiplexed, and then calculates a
likelihood. Alternatively, it is allowable that the signal
detection unit 200 is provided with one maximum-likelihood
point search unit and one replica-vector calculation unit,
25 and repeatedly performs the processes in accordance with
the number of signals to be multiplexed so that a single
system deals with a plurality of number of signals to be
multiplexed.
[0045] The configurations described in the above
30 embodiment are only examples of the content of the present
invention. The configurations can be combined with other
well-known techniques, and part of each of the
configurations can be omitted or modified without departing
22
from the scope of the present invention.
Reference Signs List
[0046] 1 communication system, 10 transmission device,
5 20 reception device, s100a, s100b, s100c, s200a, s200b,
s200c, s201a, s201b, s201c signal line, 100 precoding
unit, 30a, 30b, 30c propagation path, 200 signal
detection unit, 210 signal division unit, 220a to 220c
first maximum-likelihood point search unit, 221a to 221c
10 second maximum-likelihood point search unit, 230a to 230c
first replica-vector calculation unit, 231a to 231c second
replica-vector calculation unit, 240 likelihood
calculation unit, 300 control circuit, 300a processor,
300b memory.
15
23
WE CLAIMS
1. A reception device comprising:
a signal division unit to divide a reception signal
including a plurality of multiplexed signals respectively
5 into a real component and an imaginary component, the
multiplexed signals being obtained by multiplexing a
plurality of modulation signals by a real-number precoding
matrix, each of the modulation signals having a real
component and an imaginary component modulated
10 independently from each other;
a first maximum-likelihood point search unit to narrow
down candidate signal points, which are obtainable by a
real component of the multiplexed signal, to a first
candidate signal point by using one of the real components
15 of the reception signal;
a second maximum-likelihood point search unit to
narrow down candidate signal points, which are obtainable
by an imaginary component of the multiplexed signal, to a
second candidate signal point by using one of the imaginary
20 components of the reception signal;
a first replica-vector calculation unit to calculate a
first replica vector by using the first candidate signal
point;
a second replica-vector calculation unit to calculate
25 a second replica vector by using the second candidate
signal point; and
a likelihood calculation unit to calculate a
likelihood of the modulation signal by using the first
replica vector and the second replica vector.
30
2. The reception device according to claim 1, wherein
the first maximum-likelihood point search unit narrows
down the first candidate signal points, which are
24
obtainable by a real component of the multiplexed signal,
to the first candidate signal point by selecting a
candidate signal point located at a shortest distance from
a real component of the reception signal, and
5 the second maximum-likelihood point search unit
narrows down the second candidate signal points, which are
obtainable by an imaginary component of the multiplexed
signal, to the second candidate signal point by selecting a
candidate signal point located at a shortest distance from
10 an imaginary component of the reception signal.
3. The reception device according to claim 1 or 2,
wherein
the first replica-vector calculation unit outputs a
15 candidate signal point located at a shortest distance from
the first candidate signal point among candidate signal
points including each inverted bit of the first candidate
signal point, and
the second replica-vector calculation unit outputs a
20 candidate signal point located at a shortest distance from
the second candidate signal point among candidate signal
points including each inverted bit of the second candidate
signal point.
25 4. A communication system comprising:
a transmission device to transmit a plurality of
multiplexed signals obtained by multiplexing a plurality of
modulation signals by a real-number precoding matrix, each
of the modulation signals having a real component and an
30 imaginary component modulated independently from each
other; and
the reception device according to any one of claims 1
to 3 to receive a signal transmitted from the transmission
device.
25
5. A method for calculating a likelihood of a modulation
signal in a reception device, the method comprising:
a first step of dividing a reception signal including
5 a plurality of multiplexed signals respectively into a real
component and an imaginary component, the multiplexed
signals being obtained by multiplexing a plurality of
modulation signals by a real-number precoding matrix, each
of the modulation signals having a real component and an
10 imaginary component modulated independently from each
other;
a second step of narrowing down candidate signal
points, which are obtainable by a real component of the
multiplexed signal, to a first candidate signal point by
15 using one of the real components of the reception signal;
a third step of narrowing down candidate signal
points, which are obtainable by an imaginary component of
the multiplexed signal, to a second candidate signal point
by using one of the imaginary components of the reception
20 signal;
a fourth step of calculating a first replica vector by
using the first candidate signal point;
a fifth step of calculating a second replica vector by
using the second candidate signal point; and
25 a sixth step of calculating a likelihood of the
modulation signal by using the first replica vector and the
second replica vector.