DESCRIPTION
[Title of Invention]
SIGNAL GENERATION METHOD AND SIGNAL GENERATION DEVICE
[CROSS-REFERENCE TO RELATED APPLICATIONS]
5 This application is based on applications No. 2010-276448 filed December
10, 2010, 201 1-026422 filed February 9, 201 1, 201 1-033770 filed February 18,
20 1 1, and 201 1-05 1841 filed March 9, 201 1 in Japan, the contents of which are
hereby incorporated by reference.
[Technical Field]
10 [OOOl]
The present invention relates to a transmission device and a reception
device for communication using multiple antennas.
[Background Art]
[0002]
15 A MIMO (Multiple-Input, Multiple-Output) system is an example of a
conventional communication system using multiple antennas. In multi-antenna
communication, of which the MIMO system is typical, multiple transmission signals
are each modulated, and each modulated signal is simultaneously transmitted from a
different antenna in order to increase the transmission speed of the data.
20 [0003]
Fig. 23 illustrates a sample configuration of a transmission and reception
device having two transmit antennas and two receive antennas, and using two
transmit modulated signals (transmit streams). In the transmission device, encoded
data are interleaved, the interleaved data are modulated, and frequency conversion
25 and the like are performed to generate transmission signals, which are then
transmitted from antennas. In this case, the scheme for simultaneously transmitting
different modulated signals from different transmit antennas at the same time and on
a common frequency is a spatial multiplexing MIMO system.
1
[0004]
In this context, Patent Literature 1 suggests using a transmission device
provided with a different interleaving pattern for each transmit antenna. That is,
the transmission device fiom Fig. 23 should use two distinct interleaving patterns
5 performed by two interleavers (n, and nb). AS for the reception device, Non-Patent
Literature 1 and 2 describe improving reception quality by iteratively using soft
values for the detection scheme (by the MIMO detector of Fig. 23).
[0005]
As it happens, models of actual propagation environments in wireless
10 communications include NLOS (Non Line-Of-Sight), typified by a Rayleigh fading
environment is representative, and LOS (Line-Of-Sight), typified by a Rician fading
environment. When the transmission device transmits a single modulated signal,
and the reception device performs maximal ratio combination on the signals
received by a plurality of antennas and then demodulates and decodes the resulting
15 signals, excellent reception quality can be achieved in a LOS environment, in
particular in an environment where the Rician factor is large. The Rician factor
represents the received power of direct waves relative to the received power of
scattered waves. However, depending on the transmission system (e.g., a spatial
multiplexing MIMO system), a problem occurs in that the reception quality
20 deteriorates as the Rician factor increases (see Non-Patent Literature 3).
Figs. 24A and 24B illustrate an example of simulation results of the BER
(Bit Error Rate) characteristics (vertical axis: BER, horizontal axis: SNR
(signal-to-noise ratio) for data encoded with LDPC (low-density parity-check) codes
and transmitted over a 2x2 (two transmit antennas, two receive antennas) spatial
25 multiplexing MIMO system in a Rayleigh fading environment and in a Rician fading
environment with Rician factors of K = 3, 10, and 16 dB. Fig. 24A gives the
Max-Log approximation-based log-likelihood ratio (Max-log APP) BER
characteristics without iterative detection (see Non-Patent Literature 1 and 2), while
2
Fig. 24B gives the Max-log APP BER characteristics with iterative detection (see
Non-Patent Literature 1 and 2) (number of iterations: five). Figs. 24A and 24B
clearly indicate that, regardless of whether or not iterative detection is performed,
reception quality degrades in the spatial multiplexing MIMO system as the Rician
5 factor increases. Thus, the problem of reception quality degradation upon
stabilization of the propagation environment in the spatial multiplexing MIMO
system, which does not occur in a conventional single-modulation signal system, is
unique to the spatial multiplexing MIMO system.
[0006]
10 Broadcast or multicast communication is a service applied to various
propagation environments. The radio wave propagation environment between the
broadcaster and the receivers belonging to the users is often a LOS environment.
When using a spatial multiplexing MIMO system having the above problem for
broadcast or multicast communication, a situation may occur in which the received
15 electric field strength is high at the reception device, but in which degradation in
reception quality makes service reception difficult. In other words, in order to use
a spatial multiplexing MIMO system in broadcast or multicast communication in
both the NLOS environment and the LOS environment, a MlMO system that offers
a certain degree of reception quality is desirable.
20 [0007]
Non-Patent Literature 8 describes a scheme for selecting a codebook used in
precoding (i-e. a precoding matrix, also referred to as a precoding weight matrix)
based on feedback information fiom a communication party. However, Non-Patent
Literature 8 does not at all disclose a scheme for precoding in an environment in
25 which feedback information cannot be acquired fiom the other party, such as in the
above broadcast or multicast communication.
[OOOS]
On the other hand, Non-Patent Literature 4 discloses a scheme for switching
3
the precoding matrix over time. This scheme is applicable when no feedback
information is available. Non-Patent Literature 4 discloses using a unitary matrix
as the precoding matrix, and switching the unitary matrix at random, but does not at
all disclose a scheme applicable to degradation of reception quality in the
5 above-described LOS environment. Non-Patent Literature 4 simply recites
hopping between precoding matrices at random. Obviously, Non-Patent Literature
4 makes no mention whatsoever of a precoding method, or a structure of a precoding
matrix, for remedying degradation of reception quality in a LOS environment.
[Citation List]
10 patent Literature]
[0002]
[Patent Literature 11
International Patent Application Publication No. W020051050885
won-Patent Literature]
15 [OOlO]
won-Patent Literature 11
"Achieving near-capacity on a multiple-antenna channel" IEEE Transaction
on communications, vo1.51, no.3, pp.389-399, March 2003
[Non-Patent Literature 21
20 "Performance analysis and design optimization of LDPC-coded MIMO
OFDM systems" IEEE Trans. Signal Processing, vo1.52, no.2, pp.348-361, Feb.
2004
won-Patent Literature 31
"BER performance evaluation in 2x2 MIMO spatial multiplexing systems
25 under Rician fading channels" IEICE Trans. Fundamentals, vol.E91-A, no.10,
pp.2798-2807, Oct. 2008
[Non-Patent Literature 41
"Turbo space-time codes with time varying linear transformations" IEEE
Trans. Wireless communications, vo1.6, no.2, pp.486-493, Feb. 2007
[Non-Patent Literature 51
"Likelihood finction for QR-MLD suitable for soft-decision turbo decoding
5 and its performance" IEICE Trans. Commun., vol.ES8-B, no.1, pp.47-57, Jan. 2004
[Non-Patent Literature 61
"A tutorial on 'Parallel concatenated (Turbo) coding', 'Turbo (iterative)
decoding' and related topics" IEICE, Technical Report IT98-5 1
[Non-Patent Literature 71
10 "Advanced signal processing for PLCs: Wavelet-OFDM Proc. of IEEE
International symposium on ISPLC 2008, pp.187-192,2008
Don-Patent Literature 81
D. J. Love and R. W. Heath Jr., "Limited feedback unitary precoding for
spatial multiplexing systems" IEEE Trans. Inf. Theory, vo1.5 1, no.8, pp.2967-2976,
15 Aug. 2005
Don-Patent Literature 91
DVB Document A122, Framing structure, channel coding and modulation
for a second generation digital terrestrial television broadcasting system (DVB-T2),
June 2008
20 [Non-Patent Literature 101
L. Vangelista, N. Benvenuto, and S. Tomasin "Key technologies for
next-generation terrestrial digital television standard DVB-T2," IEEE Commun.
Magazine, vo1.47, no. 10, pp. 146-153, Oct. 2009
[Non-Patent Literature 1 I]
25 T. Ohgane, T. Nishimura, and Y. Ogawa, "Application of space division
multiplexing and those performance in a MIMO channel" IEICE Trans. Cornmun.,
vol.E88-B, no.5, pp. 1843-1 85 1, May 2005
mon-Patent Literature 121
5
R. G. Gallager "Low-density parity-check codes," IRE Trans. Inform.
Theory, IT-8, pp.21-28, 1962
won-Patent Literature 131
D. J. C. Mackay, "Good error-correcting codes based on very sparse
5 matrices," IEEE Trans. Inform. Theory, vo1.45, no.2, pp.399-431, March 1999.
won-Patent Literature 141
ETSI EN 302 307, "Second generation framing structure, channel coding
and modulation systems for broadcasting, interactive services, news gathering and
other broadband satellite applications" v. 1.1.2, June 2006
10 won-Patent Literature 151
Y.-L. Ueng, and C.-C. Cheng "A fast-convergence decoding method and
memory-efficient VLSI decoder architecture for irregular LDPC codes in the IEEE
802.16e standards" IEEE VTC-2007 Fall, pp. 1255- 1259
[Non-Patent Literature 161
15 S. M. Alamouti "A simple transmit diversity technique for wireless
communications" IEEE J. Select. Areas Commun., vol. 16, no.8, pp.145 1-1458, Oct
1998
won-Patent Literature 171
V. Tarokh, H. Jafikhani, and A. R. Calderbank "Space-time block coding
20 for wireless communications: Performance results" IEEE J. Select. Areas Commun.,
vol. 17, no.3, no.3, pp.45 1-460, March 1999
[Summary of Invention]
[Technical Problem]
[OOl 11
25 An object of the present invention is to provide a MIMO system that
improves reception quality in a LOS environment.
[Solution to Problem]
[OO 121
The present invention provides a signal generation scheme for generating,
fiom a plurality of baseband signals, a plurality of signals for transmission on a
common fiequency band and at a common time, comprising the steps of: generating
M first encoded blocks usable as a first set of bits and M second encoded blocks
5 usable as a second set of bits using a predetermined error-correcting block coding
scheme, where M is a natural number; performing a change of phase on each of a
first baseband signal sl generated fiom the first set of bits and a second baseband
signal s2 generated from the second set of bits, thus generating a first post-phase
change baseband signal sl' and a second post-phase change baseband signal s2',
10 each including M symbols; and applying weighting to the fust post-phase change
baseband signal sl' and to the second post-phase change baseband signal s2'
according to a predetermined matrix F, thus generating the plurality of signals for
transmission on the common fiequency band and at the common time as a
combination of M pairs of a first weighted signal zl and a second weighted signal 22,
15 wherein the first weighted signal zl and the second weighted signal 22 satisfj the
relation: (zl, z21T = F(slf, s2yT and the change of phase is performed on the first
baseband signal sl and the second baseband signal s2 using a phase modification
value sequentially selected fiom among N phase modification value candidates.
[00 1 31
20 Also, the present invention provides a signal generation device for
generating, fiom a plurality of baseband signals, a plurality of signals for
transmission on a common fiequency band and at a common time, comprising: an
encoder generating M first encoded blocks usable as a first set of bits and M second
encoded blocks usable as a second set of bits using a predetermined error-correcting
25 block coding scheme, where M is a natural number; a phase changer performing a
change of phase on each of a first baseband signal sl generated fiom the first set of
bits and a second baseband signal s2 generated h m the second set of bits, thus
generating a first post-phase change baseband signal sl' and a second post-phase
7
change baseband signal s2', each including M symbols; and a weighting unit
applying weighting to the first post-phase change baseband signal sl' and to the
second post-phase change baseband signal s2' according to a predetermined matrix F,
thus generating the plurality of signals for transmission on the common frequency
5 band and at the common time as a combination of M pairs of a first weighted signal
zl and a second weighted signal 22, wherein the first weighted signal zl and the
second weighted signal 22 satisfy the relation: (zl, = F(s1 ', ~ 2 'a)n~d t he change
of phase is performed on the first baseband signal sl and the second baseband signal
s2 using a phase modification value sequentially selected fiom among N phase
10 modification value candidates.
[0003]
[Advantageous Effects of Invention]
[OO 1 51
According to the above structure, the present invention provides a signal
15 generation scheme and signal generation device that remedy degradation of
reception quality in a LOS environment, thereby providing high-quality service to
LOS users during broadcast or multicast communication.
[Brief Description of Drawings]
[0004]
20 Fig. 1 illustrates an example of a transmission and reception device in a
spatial multiplexing MIMO system.
Fig. 2 illustrates a sample fkame configuration.
Fig. 3 illustrates an example of a transmission device applying a phase
changing scheme.
25 Fig. 4 illustrates another example of a transmission device applying a phase
changing scheme.
Fig. 5 illustrates another sample fiame configuration.
Fig. 6 illustrates a sample phase changing scheme.
8
Fig. 7 illustrates a sample configuration of a reception device.
Fig. 8 illustrates a sample configuration of a signal processor in the
reception device.
Fig. 9 illustrates another sample configuration of a signal processor in the
5 reception device.
Fig. 10 illustrates an iterative decoding scheme.
Fig. 1 1 illustrates sample reception conditions.
Fig. 12 illustrates a further example of a transmission device applying a
phase changing scheme.
10 Fig. 13 illustrates yet a further example of a transmission device applying a
phase changing scheme.
Fig. 14 illustrates a further sample fiame configuration.
Fig. 15 illustrates yet another sample fiame configuration.
Fig. 16 illustrates still another sample frame configuration.
Fig. 17 illustrates still yet another sample h e configuration.
Fig. 18 illustrates yet a fiuther sample frame configuration.
Figs. 19A and 19B illustrate examples of a mapping scheme.
Figs. 20A and 20B illustrate further examples of a mapping scheme.
Fig. 21 illustrates a sample configuration of a weighting unit.
Fig. 22 illustrates a sample symbol rearrangement scheme.
Fig. 23 illustrates another example of a transmission and reception device in
a spatial multiplexing MIMO system.
Figs. 24A and 24B illustrate sample BER characteristics.
Fig. 25 illustrates another sample phase changing scheme.
Fig. 26 illustrates yet another sample phase changing scheme.
Fig. 27 illustrates a hrther sample phase changing scheme.
Fig. 28 illustrates still a M e r sample phase changing scheme.
Fig. 29 illustrates still yet a hrther sample phase changing scheme.
9
Fig. 30 illustrates a sample symbol arrangement for a modulated signal
providing high received signal quality.
Fig. 31 illustrates a sample frame configuration for a modulated signal
providing high received signal quality.
5 Fig. 32 illustrates another sample symbol arrangement for a modulated
signal providing high received signal quality.
Fig. 33 illustrates yet another sample symbol arrangement for a modulated
signal providing high received signal quality.
Fig. 34 illustrates variation in numbers of symbols and slots needed per
10 coded block when block codes are used.
Fig. 35 illustrates variation in numbers of symbols and slots needed per pair
of coded blocks when block codes are used.
Fig. 36 illustrates an overall configuration of a digital broadcasting system.
Fig. 37 is a block diagram illustrating a sample receiver.
Fig. 38 illustrates multiplexed data configuration.
Fig. 39 is a schematic diagram illustrating multiplexing of encoded data into
streams.
Fig. 40 is a detailed diagram illustrating a video stream as contained in a
PES packet sequence.
20 Fig. 41 is a structural diagram of TS packets and source packets in the
multiplexed data.
Fig. 42 illustrates PMT data configuration.
Fig. 43 illustrates information as configured in the multiplexed data.
Fig. 44 illustrates the configuration of stream attribute information.
Fig. 45 illustrates the configuration of a video display and audio output
device.
Fig. 46 illustrates a sample configuration of a communications system.
Figs. 47A and 47B illustrate a variant sample symbol arrangement for a
modulated signal providing high received signal quality.
Figs. 48A and 48B illustrate another variant sample symbol arrangement for
a modulated signal providing high received signal quality.
5 Figs. 49A and 49B illustrate yet another variant sample symbol arrangement
for a modulated signal providing high received signal quality.
Figs. 50A and 50B illustrate a further variant sample symbol arrangement
for a modulated signal providing high received signal quality.
Fig. 5 1 illustrates a sample configuration of a transmission device.
Fig. 52 illustrates another sample configuration of a transmission device.
Fig. 53 illustrates a further sample configuration of a transmission device.
Fig. 54 illustrates yet a further sample configuration of a transmission
device.
Fig. 55 illustrates a baseband signal switcher.
Fig. 56 illustrates a further sample configuration of a transmission device.
Fig. 57 illustrates sample operations of a distributor.
Fig. 58 illustrates further sample operations of a distributor.
Fig. 59 illustrates a sample communications system indicating the
relationship between base stations and terminals.
20 Fig. 60 illustrates an example of transmit signal frequency allocation.
Fig. 61 illustrates another example of transmit signal frequency allocation.
Fig. 62 illustrates a sample communications system indicating the
relationship between a base station, repeaters, and terminals.
Fig. 63 illustrates an example of transmit signal frequency allocation with
25 respect to the base station.
Fig. 64 illustrates an example of transmit signal frequency allocation with
respect to the repeaters.
Fig. 65 illustrates a sample configuration of a receiver and transmitter in the
repeater.
Fig. 66 illustrates a signal data format used for transmission by the base
station.
5 Fig. 67 illustrates another sample configuration of a transmission device.
Fig. 68 illustrates another baseband signal switcher.
Fig. 69 illustrates a weighting, baseband signal switching and phase
changing scheme.
Fig. 70 illustrates a sample configuration of a transmission device using an
10 OFDM scheme.
Figs. 71A and 71B illustrate further sample fiame configurations.
Fig. 72 illustrates the numbers of slots and phase changing values
corresponding to a modulation scheme.
Fig. 73 further illustrates the numbers of slots and phase changing values
15 corresponding to a modulation scheme.
Fig. 74 illustrates the overall h e configuration of a signal transmitted by
a broadcaster using DVB-T2.
Fig. 75 illustrates two or more types of signals at the same time.
Fig. 76 illustrates a further sample configuration of a transmission device.
Fig. 77 illustrates an alternate sample frame configuration.
Fig. 78 illustrates another alternate sample fiame configuration.
Fig. 79 illustrates a further alternate sample fiame configuration.
Fig. 80 illustrates yet a hrther alternate sample frame configuration.
Fig. 81 illustrates yet another alternate sample fiame configuration.
Fig. 82 illustrates still another alternate sample came configuration.
Fig. 83 illustrates still a further alternate sample fiame configuration.
Fig. 84 further illustrates two or more types of signals at the same time.
Fig. 85 illustrates an alternate sample configuration of a transmission
device.
Fig. 86 illustrates an alternate sample configuration of a reception device.
Fig. 87 illustrates another alternate sample configuration of a reception
5 device.
Fig. 88 illustrates yet another alternate sample configuration of a reception
device.
Figs. 89A and 89B illustrate further alternate sample M e configurations.
Figs. 90A and 90B illustrate further alternate sample li-ame configurations.
10 Figs. 91A and 91B illustrate more alternate sample h e configurations.
Figs. 92A and 92B illustrate more alternate sample frame configurations.
Figs. 93A and 93B illustrate further alternate sample frame configurations.
Fig. 94 illustrates a sample frame configuration used when space-time block
codes are employed.
15 pescription of Embodiments]
[00 1 71
Embodiments of the present invention are described below with reference to
the accompanying drawings.
[Embodiment I]
20 The following describes, in detail, a transmission scheme, a transmission
device, a reception scheme, and a reception device pertaining to the present
Embodiment.
[OO 1 81
Before beginning the description proper, an outline of transmission schemes
25 and decoding schemes in a conventional spatial multiplexing MIMO system is
provided. Fig. 1 illustrates the structure of an NpN, spatial multiplexing MIMO
system. An information vector z is encoded and interleaved. The encoded bit
vector u = (ul, ... uNt) is obtained as the interleave output. Here, u, = (U,~.,. . u&
13
(where M is the number of transmitted bits per symbol). For a transmit vector s =
(s,, ... SNt), a received signal si = map(ui) is found for transmit antenna #i.
Normalizing the transmit energy, this is expressible as ~ { l s ~=l E~,/)N , (where E, is
the total energy per channel). The receive vector y = (yl, ... yN,)T is expressed in
5 Math. 1 (formula I), below.
[00 1 91
[Math. 11
(formula 1)
Here, HNwr is the channel matrix, n = (nl, . . . nN,) is the noise vector, and the
average value of ni is zero for independent and identically distributed (i.i.d) complex
Gaussian noise of variance 02. Based on the relationship between transmitted
symbols introduced into a receiver and the received symbols, the probability
15 distribution of the received vectors can be expressed as Math. 2 (formula 2), below,
for a multi-dimensional Gaussian distribution.
[002 11
[Math. 21
(formula 2)
[0022]
Here, a receiver performing iterative decoding is considered. Such a
receiver is illustrated in Fig. 1 as being made up of an outer soft-inlsofi-out decoder
and a MIMO detector. The log-likelihood ratio vector (L-value) for Fig. 1 is given
by Math. 3 (formula 3) through Math. 5 (formula 5), as follows.
[0023]
wath. 31
5 (formula 3)
[0024]
[Math. 41
(formula 4)
[0025]
[Math. 51
(formula 5)
15 [0026]
(Iterative Detection Scheme)
The following describes the MIMO signal iterative detection performed by
the NpN, spatial multiplexing MIMO system.
The log-likelihood ratio of u, is defined by Math. 6 (formula 6).
20 [0027]
[Math. 61
(formula 6)
[0028]
Through application of Bayes' theorem, Math. 6 (formula 6) can be
expressed as Math. 7 (formula 7).
5 [0029]
Math. 71
(formula 7)
P(urn=n + I )
= In + In
P(urn=n - I ) cumP(Y; ,I u )p(uI u,)
[0030]
10 Note that Urn, .I = {ulu,, = *I}. Through the approximation lnCaj - max
In a,, Math. 7 (formula 7) can be approximated as Math. 8 (formula 8). The symbol -
is herein used to signifl approximation.
[003 11
Math. 81
15 (formula 8)
In Math. 8 (formula 8), P(ulumn)a nd In P(ulumnc) an be expressed as follows.
[0033]
[Math. 91
(formula 9)
[0034]
[Math. 101
(formula 10)
10 [0035]
[Math. 111
(formula 1 1)
Note that the log-probability of the equation given in Math. 2 (formula 2)
can be expressed as Math. 12 (formula 12).
[0037]
[Math. 121
5 (formula 12)
[003 81
Accordingly, given Math. 7 (formula 7) and Math. 13 (formula 13), the
posterior L-value for the MAP or APP (a posteriori probability) can be can be
10 expressed as follows. .
[0039]
[Math. 131
(formula 13)
15 [0040]
This is hereinafter termed iterative APP decoding. Also, given Math. 8
(formula 8) and Math. 12 (formula 12), the posterior L-value for the Max-log APP
can be can be expressed as follows. .
[004 11
20 [Math. 141
(formula 14)
L(umnI Y )= Umman,x+l {y(u7y 7L (u)))- Umman,x-1 {y(u7Y ,L(u)))
(formula 15)
[0043]
5 This is hereinafter referred to as iterative Max-log APP decoding. As such,
the external information required by the iterative decoding system is obtainable by
subtracting prior input from Math. 13 (formula 13) or from Math. 14 (formula 14).
(System Model)
Fig. 23 illustrates the basic configuration of a system related to the
10 following explanations. The illustrated system is a 2x2 spatial multiplexing MIMO
system having an outer decoder for each of two streams A and B. The two outer
decoders perform identical LDPC encoding (Although the present example
considers a configuration in which the outer encoders use LDPC codes, the outer
encoders are not restricted to the use of LDPC as the error-correcting codes. The
15 example may also be realized using other error-correcting codes, such as turbo codes,
convolutional codes, or LDPC convolutional codes. Further, while the outer
encoders are presently described as individually configured for each transmit
antenna, no limitation is intended in this regard. A single outer encoder may be
used for a plurality of transmit antennas, or the number of outer encoders may be
20 greater than the number of transmit antennas. The system also has interleavers (z,
q,) for each of the streams A and B. Here, the modulation scheme is 2 h -(i-e~., ~ ~
h bits transmitted per symbol).
[0044]
The receiver performs iterative detection (iterative APP (or Max-log APP)
25 decoding) of MIMO signals, as described above. The LDPC codes are decoded
using, for example, sum-product decoding.
[0045]
19
Fig. 2 illustrates the frame configuration and describes the symbol order
after interleaving. Here, (iaja) and (ibjb)c an be expressed as follows.
[OM61
wath. 161
5 (formula 16)
[0047]
wath. 171
(formula 17)
[0048]
Here, ia and ib represent the symbol order after interleaving, ja and jb
represent the bit position in the modulation scheme (wherej ajb= 1, .. . h), na and nb
represent the interleavers of streams A and B, and Qaiea and ebrejprbesen t the data
15 order of streams A and B before interleaving. Note that Fig. 2 illustrates a situation
where ia = ib.
(Iterative Decoding)
The following describes, in detail, the sum-product decoding used in
decoding the LDPC codes and the MIMO signal iterative detection algorithm, both
20 used by the receiver.
[0049]
Sum-Product Decoding
A two-dimensional MxN matrix H = {Hm) is used as the check matrix for
LDPC codes subject to decoding. For the set [l,N] = (1, 2 ... N), the partial sets
25 A(m) and B(n) are defined as follows.
[OOSO]
20
Math. 181
(formula 18)
A(m)={n:H rnn =I}
[005 11
5 [Math. 191
(formula 19)
B(n)={m:Hrnn =1}
[0052]
Here, A(m) signifies the set of column indices equal to 1 for row m of check
10 matrix H, while B(n) signifies the set of row indices equal to 1 for row n of check
matrix H. The sum-product decoding algorithm is as follows.
Step A- 1 (Initialization): For all pairs (m,n) satisfling H,, = 1, set the prior
log ratio P, = 1. Set the loop variable (number of iterations) 1,- = 1, and set the
maximum number of loops
15 Step A-2 (Processing): For all pairs (m,n) satisfling H,, = 1 in the order m = 1,2,
. . . M , update the extrinsic value log ratio a, using the following update formula.
[0053]
[Math. 201
(formula 20)
rnn
nl.sA(rn)\n n'~A(rn)\n
[0054]
[Math. 211
(formula 21)
[Math. 221
(formula 22)
exp(x) + 1
5 f (x) - In exp(x) - 1
where f is the Gallager function. h, can then be computed as follows.
Step A-3 (Column Operations): For all pairs (m,n) satisfling H, = 1 in the order n
= 1, 2, ... N , update the extrinsic value log ratio P, using the following update
10 formula.
[0057]
[Math. 231
(formula 23)
Step A-4 (Log-likelihood Ratio Calculation): For nE[l,N], the log-likelihood
ratio L, is computed as follows.
[Math. 241
20 (formula 24)
Step A-5 (Iteration Count): If 1,- < I, then 1,- is incremented and the
process returns to step A-2. Sum-product decoding ends when l,, = l,,,.
[006 11
The above describes one iteration of sum-product decoding operations.
Afterward, MIMO signal iterative detection is performed. The variables m, n, a,,
p, h,, and L, used in the above explanation of sum-product decoding operations
b 5 are expressed as m, n, a"-, pam A- and L, for stream A and as mb nb a ,hb,
pbmbnb, and Lnb for stream B.
(MIMO Signal Iterative Detection)
The following describes the calculation of h, for MIMO signal iterative
detection.
10 [0062]
The following formula is derivable fi-om Math. 1 (formula 1).
[0063]
Math. 251
(formula 25)
[0064]
Given the fkame configuration illustrated in Fig. 2, the following functions
are derivable fi-om Math. 16 (formula 16) and Math. 17 (formula 17).
[0065]
20 [Math. 261
(formula 26)
[0066]
[Math. 271
25 (formula 27)
[0067]
where n,nb E [l,N]. For iteration k of MIMO signal iterative detection, the
variables A- Ln, Lb, and Lnb are expressed as hknm Lb L,nb and Lk,nb.
5 [0068]
Step B-1 (Initial Detection; k = 0)
For initial wave detection, h,, and b,nabre calculated as follows.
For iterative APP decoding:
[0069]
10 wath. 281
(formula 28)
[0070]
For iterative Max-log APP decoding:
15 [0071]
[Math. 291
(formula 29)
[0072]
20 [Math. 301
(formula 30)
[0073]
where X = a,b. Next, the iteration count for the MIMO signal iterative
detection is set to I,,, = 0, with the maximum iteration count being I,-.
5 Step B-2 (Iterative Detection; Iteration k): When the iteration count is k,
Math. 1 1 (formula 1 I), Math. 13 (formula 13) through Math. 15 (formula 15), Math.
16 (formula 16), and Math. 17 (formula 17) can be expressed as Math. 31 (formula
31) through Math. 34 (formula 34), below. Note that (X,Y) = (a,b)(b,a).
For iterative APP decoding:
10 [0074]
wath. 3 11
(formula 3 1)
(formula 32)
[0076]
For iterative Max-log APP decoding:
20 [0077]
Wath. 331
25
(formula 33)
[0078]
[Math. 341
5 (formula 34)
[0079]
Step B-3 (Iteration Count and Codeword Estimation) If ldmo <
then I,, is incremented and the process returns to step B-2. When =
10 an estimated codeword is found, as follows.
[0080]
Wath. 351
(formula 35)
15 [0081]
where X = a,b.
Fig. 3 shows a sample configuration of a transmission device 300 pertaining
to the present Embodiment. An encoder 302A takes information (data) 301A and a
M e configuration signal 3 13 as input (which includes the error-correction scheme,
20 coding rate, block length, and other information used by the encoder 302A in
error-correction coding of the data, such that the scheme designated by the frame
configuration signal 313 is used. The error-correction scheme may be switched).
In accordance with the frame configuration signal 313, the encoder 302A performs
error-correction coding, such as convolutional encoding, LDPC encoding, turbo
encoding or similar, and outputs encoded data 303A.
[0082]
An interleaver 304A takes the encoded data 303A and the frame
configuration signal 313 as input, performs interleaving, i.e., rearranges the order
thereof, and then outputs interleaved data 305A. (Depending on the frame
configuration signal 3 13, the interleaving scheme may be switched.)
A mapper 306A takes the interleaved data 305A and the frame
configuration signal 313 as input and performs modulation, such as QPSK
(Quadrature Phase Shift Keying), 16-QAM (1 6-Quadradature Amplitude
Modulation), or 64-QAM (64-Quadradture Amplitude Modulation) thereon, then
outputs a baseband signal 307A. (Depending on the frame configuration signal 3 13,
the modulation scheme may be switched.)
Figs. 19A and 19B illustrate an example of a QPSK modulation mapping
scheme for a baseband signal made up of an in-phase component I and a quadrature
component Q in the IQ plane. For example, as shown in Fig. 19A, when the input
data are 00, then the output is I = 1 .O, Q = 1.0. Similarly, when the input data are
01, the output is I = -1.0, Q = 1.0, and so on. Fig. 19B illustrates an example of a
QPSK modulation mapping scheme in the IQ plane differing from Fig. 19A in that
the signal points of Fig. 19A have been rotated about the origin to obtain the signal
points of Fig. 19B. Non-Patent Literature 9 and Non-Patent Literature 10 describe
such a constellation rotation scheme. Alternatively, the Cyclic Q Delay described
in Non-Patent Literature 9 and Non-Patent Literature 10 may also be adopted. An
alternate example, distinct from Figs. 19A and 19B, is shown in Figs. 20A and 20B,
which illustrate a signal point layout for 16-QAM in the IQ plane. The example of
Fig. 20A corresponds to Fig. 19A, while that of Fig. 20B corresponds to Fig. 19B.
[0083]
An encoder 302B takes information (data) 301B and the fiame
configuration signal 313 as input (which includes the error-correction scheme,
coding rate, block length, and other information used by the encoder 302A in
error-correction coding of the data, such that the scheme designated by the £kame
5 configuration signal 313 is used. The error-correction scheme may be switched).
In accordance with the fiame configuration signal 313, the encoder 302B performs
error-correction coding, such as convolutional encoding, LDPC encoding, turbo
encoding or similar, and outputs encoded data 303B.
[0084]
10 An interleaver 304B takes the encoded data 303B and the frame
configuration signal 313 as input, performs interleaving, i.e., rearranges the order
thereof, and outputs interleaved data 305B. (Depending on the b m e configuration
signal 3 13, the interleaving scheme may be switched.)
A mapper 306B takes the interleaved data 305B and the frame configuration
15 signal 3 13 as input and performs modulation, such as QPSK, 16-QAM, or 64-QAM
thereon, then outputs a baseband signal 307B. (Depending on the fiame
configuration signal 3 13, the modulation scheme may be switched.)
A signal processing scheme information generator 314 takes the b e
configuration signal 3 13 as input and accordingly outputs signal processing scheme
20 information 315. The signal processing scheme information 3 15 designates the
fned precoding matrix to be used, and includes information on the pattern of phase
changes used for changing the phase.
[0085]
A weighting unit 308A takes baseband signal 307A, baseband signal 307B,
and the signal processing scheme information 315 as input and, in accordance with
the signal processing scheme information 315, performs weighting on the baseband
signals 307A and 307I3, then outputs a weighted signal 309A. The weighting
scheme is described in detail, later.
28
[0086]
A wireless unit 310A takes weighted signal 309A as input and performs
processing such as quadrature modulation, band limitation, frequency conversion,
amplification, and so on, then outputs transmit signal 31 1A. Transmit signal 31 1A
5 is then output as radio waves by an antenna 3 12A.
[0087]
A weighting unit 308B takes baseband signal 307A, baseband signal 307B,
and the signal processing scheme information 3 15 as input and, in accordance with
the signal processing scheme information 315, performs weighting on the baseband
10 signals 307A and 307B, then outputs weighted signal 3 16B.
[OOSS]
Fig. 21 illustrates the configuration of the weighting units 308A and 308B.
The area of Fig. 21 enclosed in the dashed line represents one of the weighting units.
Baseband signal 307A is multiplied by wll to obtain wll-sl(t), and multiplied by
15 w21 to obtain ~ 2 1 .(t~). 1 Similarly, baseband signal 307B is multiplied by w12 to
obtain w12-s2(t), and multiplied by w22 to obtain w22.s2(t). Next, zl(t) =
wl l.sl(t) + w12-s2(t) and z2(t) = w21*sl(t) + w22-s22(t) are obtained. Here, as
explained above, sl(t) and s2(t) are baseband signals modulated according to a
modulation scheme such as BPSK (Binary Phase Shift Keying), QPSK, 8-PSK
20 (8-Phase Shift Keying), 16-QAM, 32-QAM (32-Quadrature Amplitude Modulation),
64-QAM, 256-QAM 16-APSK (16-Amplitude Phase Shift Keying) and so on.
[0089]
Both weighting units perform weighting using a fxed precoding matrix.
The precoding matrix uses, for example, the scheme of Math. 36 (formula 36), and
25 satisfies the conditions of Math. 37 (formula 37) or Math. 38 (formula 38), all found
below. However, this is only an example. The value of a is not restricted to
Math. 37 (formula 37) and Math. 38 (formula 38), and may take on other values, e.g.,
a= 1.
29
Here, the precoding matrix is:
5 (formula 36)
In Math. 36 (formula 36), above, a may be given by:
[0093]
10 [Math. 371
(formula 37)
Alternatively, in Math. 36 (formula 36), above, a may be given by:
15 [0095]
[Math. 381
(formula 38)
20 The precoding matrix is not restricted to that of Math. 36 (formula 36), but
may also be as indicated by Math. 39 (formula 39).
[0097]
[Math. 391
(formula 39)
[0098]
In Math. 39 (forinula 39), let a = ~ d " "b = c = cdS2',a nd d = ~ e " ~ ~ .
Further, one of a, b, c, and d may be zero. For example, the following
5 configurations are possible: (1) a may be zero while b, c, and d are non-zero, (2) b
may be zero while a, c, and d are non-zero, (3) c may be zero while a, b, and d are
non-zero, or (4) d may be zero while a, b, and c are non-zero.
[0099]
When any of the modulation scheme, error-correcting codes, and the coding
10 rate thereof are changed, the precoding matrix may also be set, changed, and fixed
for use.
[O 1001
A phase changer 317B takes weighted signal 316B and the signal
processing scheme information 315 as input, then regularly changes the phase of the
15 signal 316B for output. This regular change is a change of phase performed
according to a predetermined phase changing pattern having a predetermined period
(cycle) (e.g., every n symbols (n being an integer, n 2 1) or at a predetermined
interval). The details of the phase changing pattern are explained below, in
Embodiment 4.
20 [OlOl]
Wireless unit 310B takes post-phase change signal 309B as input and
performs processing such as quadrature modulation, band limitation, frequency
conversion, amplification, and so on, then outputs transmit signal 31 1B. Transmit
signal 3 11B is then output as radio waves by an antenna 3 12B.
25 [0102]
Fig. 4 illustrates a sample configuration of a transmission device 400 that
differs from that of Fig. 3. The points of difference of Fig. 4 from Fig. 3 are
described next.
An encoder 402 takes information (data) 401 and the frame configuration
5 signal 3 13 as input, and, in accordance with the frame configuration signal 3 13,
performs error-correction coding and outputs encoded data 402.
[0 1031
A distributor 404 takes the encoded data 403 as input, performs distribution
thereof, and outputs data 405A and data 405B. Although Fig. 4 illustrates only one
10 encoder, the number of encoders is not limited as such. The present invention may
also be realized using m encoders (m being an integer, m 2 1) such that the
distributor divides the encoded data created by each encoder into two groups for
distribution.
[0 1 041
15 Fig. 5 illustrates an example of a frame configuration in the time domain for
a transmission device according to the present Embodiment. Symbol 500-1 is for
notifling the reception device of the transmission scheme. For example, symbol
500-1 conveys information such as the error-correction scheme used for
transmitting data symbols, the coding rate thereof, and the modulation scheme used
20 for transmitting data symbols.
[0 1051
Symbol 501-1 is for estimating channel fluctuations for modulated signal
zl(t) (where t is time) transmitted by the transmission device. Symbol 502-1 is a
data symbol transmitted by modulated signal zl(t) as symbol number u (in the time
25 domain). Symbol 503-1 is a data symbol transmitted by modulated signal zl(t) as
symbol number u+l.
[0 1061
Symbol 5012 is for estimating channel fluctuations for modulated signal
z2(t) (where t is time) transmitted by the transmission device. Symbol 502-2 is a
data symbol transmitted by modulated signal z2(t) as symbol number u (in the time
domain). Symbol 503-2 is a data symbol transmitted by modulated signal zl(t) as
5 symbol number u+ 1.
[0 1071
Here, the symbols of zl(t) and of z2(t) having the same time (identical
timing) are transmitted fiom the transmit antenna using the same (sharedcommon)
fiequency .
10 [0108]
The following describes the relationships between the modulated signals
zl(t) and z2(t) transmitted by the transmission device and the received signals rl(t)
and r2(t) received by the reception device.
In Fig. 5, 504#1 and 504#2 indicate transmit antennas of the transmission
15 device, while 505#1 and 505#2 indicate receive antennas of the reception device.
The transmission device transmits modulated signal zl(t) fiom transmit antenna
504#1 and transmits modulated signal z2(t) fiom transmit antenna 504#2. Here,
the modulated signals zl(t) and z2(t) are assumed to occupy the same
(sharedcommon) fiequency (bandwidth). The channel fluctuations in the transmit
20 antennas of the transmission device and the antennas of the reception device are
hl l(t), h12(t), hZl(t), and hZ2(t), respectively. Assuming that receive antenna 505# 1
of the reception device receives received signal rl(t) and that receive antenna 505#2
of the reception device receives received signal r2(t), the following relationship
holds.
25 [0109]
[Math. 401
(formula 40)
Fig. 6 pertains to the weighting scheme (precoding scheme) and the phase
changing scheme of the present Embodiment. A weighting unit 600 is a combined
5 version of the weighting units 308A and 308B from Fig. 3. As shown, stream sl(t)
and stream s2(t) correspond to the baseband signals 307A and 307B of Fig. 3. That
is, the streams sl(t) and s2(t) are baseband signals made up of an in-phase
component I and a quadrature component Q conforming to mapping by a
modulation scheme such as QPSK, 16-QAM, and 64-QAM. As indicated by the
10 h e configuration of Fig. 6, stream sl (t) is represented as sl (u) at symbol number
u, as sl(u+l) at symbol number u+l, and so forth. Similarly, stream s2(t) is
represented as s2(u) at symbol number u, as s2(u+l) at symbol number u+l, and so
forth. The weighting unit 600 takes the baseband signals 307A (sl(t)) and 307B
(s2(t)) as well as the signal processing scheme information 3 15 from Fig. 3 as input,
15 performs weighting in accordance with the signal processing scheme information
315, and outputs the weighted signals 309A (zl(t)) and 316B(z2'(t)) from Fig. 3.
The phase changer 317B changes the phase of weighted signal 316B(z2'(t)) and
outputs post-phase change signal 309B(z2(t)).
[Olll]
Here, given vector Wl = (wl l,w12) from the first row of the fixed
precoding matrix F, zl(t) is expressible as Math. 41 (formula 41), below.
[0112]
wath. 411
(formula 4 1)
25 zl(t) = wl x (sl(t),~ 2 ( t ) ) ~
[0113]
Similarly, given vector W2 = (w21,w22) fiom the second row of the fixed
preceding matrix F, and letting the phase changing formula applied by the phase
changer by y(t), then z2(t) is expressible as Math. 42 (formula 42), below.
[0114]
5 [Math. 421
(formula 42)
z2(t) = y(t) x w2 x (sl(t),~ 2 ( t ) ) ~
[0115]
Here, y(t) is a phase changing formula following a predetermined scheme.
10 For example, given a period (cycle) of four and time u, the phase changing formula
is expressible as Math. 43 (formula 43), below.
[0116]
[Math. 431
(formula 43)
[0117]
Similarly, the phase changing formula for time u+l may be, for example, as
given by Math. 44 (formula 44).
[0118]
20 [Math. 441
(formula 44)
[0119]
That is, the phase changing formula for time u+k is expressible as Math. 45
25 (formula 45).
[O 1 201
[Math. 451
(formula 45)
. kz 1-
[0121]
5 Note that Math. 43 (formula 43) through Math. 45 (formula 45) are given
only as an example of regular phase changing.
The regular change of phase is not restricted to a period (cycle) of four.
Improved reception capabilities (the error-correction capabilities, to be exact) may
potentially be promoted in the reception device by increasing the period (cycle)
10 number (this does not mean that a greater period (cycle) is better, though avoiding
small numbers such as two is likely ideal).
[O 1221
Furthermore, although Math. 43 (formula 43) through Math. 45 (formula
45), above, represent a configuration in which a change in phase is carried out
15 through rotation by consecutive predetermined phases (in the above formula, every
n./2), the change in phase need not be rotation by a constant amount, but may also be
random. For example, in accordance with the predetermined period (cycle) of y(t),
the phase may be changed through sequential multiplication as shown in Math. 46
(formula 46) and Math. 47 (formula 47). The key point of regular phase changing
20 is that the phase of the modulated signal is regularly changed. The degree of phase
change is preferably as even as possible, such as from -n. radians to n. radians.
However, given that this describes a distribution, random changes are also possible.
[O 1231
wath. 461
25 (formula 46)
[0 1241
[Math. 471
(formula 47)
3 . 5 ~ .7n
j-n J- J- +e -+e +e 4
[0 1251
As such, the weighting unit 600 of Fig. 6 performs precoding using fixed,
predetermined precoding weights, and the phase changer 3 17B changes the phase of
the signal input thereto while regularly varying the phase changing degree.
When a specialized precoding matrix is used in a LOS environment, the
reception quality is likely to improve tremendously. However, depending on the
direct wave conditions, the phase and amplitude components of the direct wave may
greatly differ from the specialized precoding matrix, upon reception. The LOS
15 environment has certain rules. Thus, data reception quality is tremendously
improved through a regular change applied to a transmit signal that obeys those
rules. The present invention offers a signal processing scheme for improvements
in the LOS environment.
[0 1 271
20 Fig. 7 illustrates a sample configuration of a reception device 700 pertaining
to the present embodiment. Wireless unit 703-X receives, as input, received signal
702-X received by antenna 701-X, performs processing such as frequency
conversion, quadrature demodulation, and the like, and outputs baseband signal
704-X.
[0 1281
Channel fluctuation estimator 705-1 for modulated signal zl transmitted by
5 the transmission device takes baseband signal 704-X as input, extracts reference
symbol 501-1 for channel estimation from Fig. 5, estimates the value of hll from
Math. 40 (formula 40), and outputs channel estimation signal 706-1.
[0 1291
Channel fluctuation estimator 705-2 for modulated signal 22 transmitted by
10 the transmission device takes baseband signal 704-X as input, extracts reference
symbol 501-2 for channel estimation from Fig. 5, estimates the value of h12 from
Math. 40 (formula 40), and outputs channel estimation signal 706-2.
[0130]
Wire1ess"unit 703-Y receives, as input, received signal 702-Y received by
15 antenna 701-X, performs processing such as frequency conversion, quadrature
demodulation, and the like, and outputs baseband signal 704-Y.
Channel fluctuation estimator 707-1 for modulated signal zl transmitted by
the transmission device takes baseband signal 704-Y as input, extracts reference
symbol 501-1 for channel estimation from Fig. 5, estimates the value of hzl from
20 Math. 40 (formula 40), and outputs channel estimation signal 708-1.
[0131]
Channel fluctuation estimator 707-2 for modulated signal 22 transmitted by
the transmission device takes baseband signal 70- as input, extracts reference
symbol 501-2 for channel estimation from Fig. 5, estimates the value of hz2 from
25 Math. 40 (formula 40), and outputs channel estimation signal 708-2.
[0132]
A control information decoder 709 receives baseband signal 704-X and
baseband signal 704-Y as input, detects symbol 500-1 that indicates the
38
transmission scheme fiom Fig. 5, and outputs a transmission scheme information
signal 7 10 for the transmission device.
[0133]
A signal processor 71 1 takes the baseband signals 704-X and 704-Y, the
5 channel estimation signals 706 -1, 706-2, 708-1, and 708-2, and the transmission
scheme information signal 710 as input, performs detection and decoding, and then
outputs received data 712-1 and 712-2.
[0134]
Next, the operations of the signal processor 71 1 fiom Fig. 7 are described in
10 detail. Fig. 8 illustrates a sample configuration of the signal processor 711
pertaining to the present embodiment. As shown, the signal processor 711 is
primarily made up of an inner MIMO detector, soft-inlsofi-out decoders, and a
coefficient generator. Non-Patent Literature 2 and Non-Patent Literature 3
describe a scheme of iterative decoding using this structure. The MIMO system
15 described in Non-Patent Literature 2 and Non-Patent Literature 3 is a spatial
multiplexing MIMO system, while the present Embodiment differs fiom Non-Patent
Literature 2 and Non-Patent Literature 3 in describing a MIMO system that regularly
changes the phase over time while using the same precoding matrix. Taking the
(channel) matrix H(t) of Math. 36 (formula 36), then by letting the precoding weight
20 matrix fiom Fig. 6 be F (here, a fixed precoding matrix remaining unchanged for a
given received signal) and letting the phase changing formula used by the phase
changer from Fig. 6 be Y(t) (here, Y(t) changes over time t), then the receive vector
R(t) = (rl(t),r2(t)lT and the stream vector S(t) = (sl(t),s2(t)lT the following fbnction
is derived:
25 [0135]
Math. 481
(formula 48)
where
[0136]
5 Here, the reception device may use the decoding schemes of Non-Patent
Literature 2 and 3 on R(t) by computing H(t)xY(t)xF.
Accordingly, the coefficient generator 819 from Fig. 8 takes a transmission
scheme information signal 818 (corresponding to 710 from Fig. 7) indicated by the
transmission device (information for specifying the fixed precoding matrix in use
10 and the phase changing pattern used when the phase is changed) and outputs a signal
processing scheme information signal 820.
[0137]
The inner MIMO detector 803 takes the signal processing scheme
information signal as input and performs iterative detection and decoding using the
15 signal and the relationship thereof to Math. 48 (formula 48). The operations
thereof are described below.
[0138]
The processing unit illustrated in Fig. 8 uses a processing scheme, as
illustrated by Fig. 10, to perform iterative decoding (iterative detection). First,
20 detection of one codeword (or one h e ) of modulated signal (stream) sl and of
one codeword (or one h e ) of modulated signal (stream) s2 is performed. As a
result, the soft-inlsoft-out decoder obtains the log-likelihood ratio of each bit of the
codeword (or W e ) of modulated signal (stream) sl and of the codeword (or h e )
of modulated signal (stream) s2. Next, the log-likelihood ratio is used to perform a
25 second round of detection and decoding. These operations are performed multiple
times (these operations are hereinafter referred to as iterative decoding (iterative
detection)). The following explanations center on the creation scheme of the
40
log-likelihood ratio of a symbol at a specific time within one frame.
[0139]
In Fig. 8, a memory 815 takes baseband signal 801X (corresponding to
baseband signal 704-X from Fig. 7), channel estimation signal group 802X
5 (corresponding to channel estimation signals 706-1 and 706-2 from Fig. 7),
baseband signal 801Y (corresponding to baseband signal 704-Y from Fig. 7), and
channel estimation signal group 802Y (corresponding to channel estimation signals
708-1 and 708-2 from Fig. 7) as input, executes (computes) H(t)xY(t)xF from Math.
48 (formula 48) in order to perform iterative decoding (iterative detection) and
10 stores the resulting matrix as a transformed channel signal group. The memory 815
then outputs the above-described signals as needed, specifically as baseband signal
8 16X, transformed channel estimation signal group 8 1 7X, baseband signal 8 16Y,
and transformed channel estimation signal group 81 7Y.
[0 1 401
15 Subsequent operations are described separately for initial detection and for
iterative decoding (iterative detection).
(Initial Detection)
The inner MIMO detector 803 takes baseband signal 801X, channel
estimation signal group 802X, baseband signal 801Y, and channel estimation signal
20 group 802Y as input. Here, the modulation scheme for modulated signal (stream)
sl and modulated signal (stream) s2 is taken to bel6-QAM.
[0141]
The inner MIMO detector 803 first computes H(t)xY(t)xF from the channel
estimation signal groups 802X and 802Y, thus calculating a candidate signal point
25 corresponding to baseband signal 801X. Fig. 11 represents such a calculation. In
Fig. 11, each black dot is a candidate signal point in the IQ plane. Given that the
modulation scheme is 16-QAM, 256 candidate signal points exist. (However, Fig.
11 is only a representation and does not indicate all 256 candidate signal points.)
41
Letting the four bits transmitted in modulated signal sl be bO, bl, b2, and b3 and the
four bits transmitted in modulated signal s2 be b4, b5, b6, and b7, candidate signal
points corresponding to (bO, bl, b2, b3, b4, b5, b6, b7) are found in Fig. 11. The
Euclidean squared distance between each candidate signal point and each received
5 signal point 1101 (corresponding to baseband signal 801X) is then computed. The
Euclidian squared distance between each point is divided by the noise variance 02.
Accordingly, Ex(bO, bl, b2, b3, b4, b5, b6, b7) is calculated. That is, Ex is the
Euclidian squared distance between a candidate signal point corresponding to (bO,
bl, b2, b3, b4, b5, b6, b7) and a received signal point, divided by the noise variance.
10 Here, each of the baseband signals and the modulated signals sl and s2 is a complex
signal.
[0 1421
Similarly, the inner MIMO detector 803 computes H(t)xY(t)xF from the
channel estimation signal groups 802X and 802Y, calculates candidate signal points
15 corresponding to baseband signal 801Y, computes the Euclidean squared distance
between each of the 'candidate signal points and the received signal points
(corresponding to baseband signal 801Y), and divides the Euclidean squared
distance by the noise variance 02. Accordingly, Ey(b0, bl, b2, b3, b4, b5, b6, b7)
is calculated. That is, Ey is the Euclidian squared distance between a candidate
20 signal point corresponding to (bO, bl, b2, b3, b4, b5, b6, b7) and a received signal
point, divided by the noise variance.
[0 1431
Next, Ex(bO, bl, b2, b3, b4, b5, b6, b7) + Ey(bO, bl, b2, b3, b4, b5, b6, b7)
= E(b0, bl, b2, b3, b4, b5, b6, b7) is computed.
25 [0144]
The inner MIMO detector 803 outputs E(b0, bl, b2, b3, b4, b5, b6, b7) as a
signal 804.
Log-likelihood calculator 805A takes the signal 804 as input, calculates the
log-likelihood of bits bO, bl, b2, and b3, and outputs log-likelihood signal 806A.
Note that this log-likelihood calculation produces the log-likelihood of a bit being 1
and the log-likelihood of a bit being 0. The calculation scheme is as shown in
5 Math. 28 (formula 28), Math. 29 (formula 29), and Math. 30 (formula 30), and the
details are given by Non-Patent Literature 2 and 3.
[0145]
Similarly, log-likelihood calculator 805A takes the signal 804 as input,
calculates the log-likelihood of bits bO, bl, b2, and b3, and outputs log-likelihood
10 signal 806B. A deinterleaver (807A) takes log-likelihood signal 806A as input,
performs deinterleaving corresponding to that of the interleaver (the interleaver
(304A) fiom Fig. 3), and outputs deinterleaved log-likelihood signal 808A.
[0 1461
Similarly, a deinterleaver (807B) takes log-likelihood signal 806B as input,
15 performs deinterleaving corresponding to that of the interleaver (the interleaver
(304B) fiom Fig. 3), and outputs deinterleaved log-likelihood signal 808B.
[0 1471
Log-likelihood ratio calculator 809A takes deinterleaved log-likelihood
signal 808A as input, calculates the log-likelihood ratio of the bits encoded by
20 encoder 302A fiom Fig. 3, and outputs log-likelihood ratio signal 810A.
[0 1481
Similarly, log-likelihood ratio calculator 809B takes deinterleaved
log-likelihood signal 808B as input, calculates the log-likelihood ratio of the bits
encoded by encoder 302B fiom Fig. 3, and outputs log-likelihood ratio signal 810B.
25 [0149]
Soft-inlsoft-out decoder 8 1 1 A takes log-likelihood ratio signal 8 10A as
input, performs decoding, and outputs decoded log-likelihood ratio 812A.
Similarly, soft-idsoft-out decoder 81 1B takes log-likelihood ratio signal
810B as input, performs decoding, and outputs decoded log-likelihood ratio 812B.
[0150]
(Iterative Decoding (Iterative Detection), k Iterations)
5 The interleaver (8 13A) takes the k- 1 th decoded log-likelihood ratio 8 12A
decoded by the soft-idsoft-out decoder as input, performs interleaving, and outputs
interleaved log-likelihood ratio 814A. Here, the interleaving pattern used by the
interleaver (813A) is identical to that of the interleaver (304A) from Fig. 3.
[0151]
10 Another interleaver (8 13B) takes the k-lth decoded log-likelihood ratio
812B decoded by the soft-idsoft-out decoder as input, performs interleaving, and
outputs interleaved log-likelihood ratio 814B. Here, the interleaving pattern used
by the other interleaver (813B) is identical to that of another interleaver (304B) from
Fig. 3.
15 [0152]
The inner MIMO detector 803 takes baseband signal 816X, transformed
channel estimation signal group 81 7X, baseband signal 816Y, transformed channel
estimation signal group 8 1711, interleaved log-likelihood ratio 8 14A, and interleaved
log-likelihood ratio 8 14B as input. Here, baseband signal 8 16X, transformed
20 channel estimation signal group 817X, baseband signal 816Y, and transformed
channel estimation signal group 817Y are used instead of baseband signal 801X,
channel estimation signal group 802X, baseband signal 801Y, and channel
estimation signal group 802Y because the latter cause delays due to the iterative
decoding.
25 [0153]
The iterative decoding operations of the inner MIMO detector 803 differ
from the initial detection operations thereof in that the interleaved log-likelihood
ratios 814A and 814B are used in signal processing for the former. The inner
44
MIMO detector 803 first calculates E(b0, bl, b2, b3, b4, b5, b6, b7) in the same
manner as for initial detection. In addition, the coefficients corresponding to Math.
11 (formula 11) and Math. 32 (formula 32) are computed fiom the interleaved
log-likelihood ratios 814A and 814B. The value of E(b0, bl, b2, b3, b4, b5, b6,
5 b7) is corrected using the coefficients so calculated to obtain E'(b0, bl, b2, b3, b4,
b5, b6, b7), which is output as the signal 804.
[0 1 541
Log-likelihood calculator 805A takes the signal 804 as input, calculates the
log-likelihood of bits bO, bl, b2, and b3, and outputs the log-likelihood signal 806A.
10 Note that this log-likelihood calculation produces the log-likelihood of a bit being 1
and the log-likelihood of a bit being 0. The calculation scheme is as shown in
Math. 31 (formula 31) through Math. 35 (formula 35), and the details are given by
Non-Patent Literature 2 and 3.
[0155]
15 Similarly, log-likelihood calculator 805B takes the signal 804 as input,
calculates the log-likelihood of bits b4, b5, b6, and b7, and outputs the
log-likelihood signal 806A. Operations performed by the deinterleaver onwards
are similar to those performed for initial detection.
[0156]
20 While Fig. 8 illustrates the configuration of the signal processor when
performing iterative detection, this structure is not absolutely necessary as good
reception improvements are obtainable by iterative detection alone. As long as the
components needed for iterative detection are present, the configuration need not
include the interleavers 813A and 813B. In such a case, the inner MIMO detector
25 803 does not perform iterative detection.
[0157]
The key point for the present Embodiment is the calculation of H(t)xY(t)xF.
As shown in Non-Patent Literature 5 and the like, QR decomposition may also be
used to perform initial detection and iterative detection.
Also, as indicated by Non-Patent Literature 11, MMSE (Minimum
5 Mean-Square Error) and ZF (Zero-Forcing) linear operations may be performed
based on H(t)xY(t)xF when performing initial detection.
[0158]
Fig. 9 illustrates the configuration of a signal processor, unlike that of Fig. 8,
that serves as the signal processor for modulated signals transmitted by the
10 transmission device fiom Fig. 4. The point of difference from Fig. 8 is the number
of soft-inlsoft-out decoders. A soft-inlsoft-out decoder 901 takes the log-likelihood
ratio signals 810A and 81 0B as input, performs decoding, and outputs a decoded
log-likelihood ratio 902. A distributor 903 takes the decoded log-likelihood ratio
902 as input for distribution. Otherwise, the operations are identical to those
15 explained for Fig. 8.
[0159]
As described above, when a transmission device according to the present
Embodiment using a MIMO system transmits a plurality of modulated signals from
a plurality of antennas, changing the phase over time while multiplying by the
20 precoding matrix so as to regularly change the phase results in improvements to data
reception quality for a reception device in a LOS environment where direct waves
are dominant, in contrast to a conventional spatial multiplexing MIMO system.
[0 1601
In the present Embodiment, and particularly in the configuration of the
25 reception device, the number of antennas is limited and explanations are given
accordingly. However, the Embodiment may also be applied to a greater number
of antennas. In other words, the number of antennas in the reception device does
not affect the operations or advantageous effects of the present Embodiment.
46
[0161]
Also, although LDPC codes are described as a particular example, the
present Embodiment is not limited in this manner. Furthermore, the decoding
scheme is not limited to the sum-product decoding example given for the
5 soft-idsoft-out decoder. Other soft-idsofi-out decoding schemes, such as the
BCJR algorithm, SOVA, and the Max-Log-Map algorithm may also be used.
Details are provided in Non-Patent Literature 6.
[0 1 621
In addition, although the present Embodiment is described using a
10 single-carrier scheme, no limitation is intended in this regard. The present
Embodiment is also applicable to multi-carrier transmission. Accordingly, the
present Embodiment may also be realized using, for example, spread-spectrum
communications, OFDM' (Orthogonal Frequency-Division Multiplexing),
SC-FDMA (Single Carrier Frequency-Division Multiple Access), SC-OFDM
1 5 (Single Carrier Orthogonal Frequency-Division Multiplexing), wavelet OFDM as
described in Non-Patent Literature 7, and so on. Furthermore, in the present
Embodiment, symbols other than data symbols, such as pilot symbols (preamble,
unique word, etc) or symbols transmitting control information, may be arranged
within the h e in any manner.
20 [0163]
The following describes an example in which OFDM is used as a
multi-carrier scheme.
Fig. 12 illustrates the configuration of a transmission device using OFDM.
In Fig. 12, components operating in the manner described for Fig. 3 use identical
25 reference numbers.
[0 1641
OFDM-related processor 1201A takes weighted signal 309A as input,
performs OFDM-related processing thereon, and outputs transmit signal 1202A.
47
Similarly, OFDM-related processor 1201B takes post-phase change 309B as input,
performs OFDM-related processing thereon, and outputs transmit signal 1202A
[0165]
Fig. 13 illustrates a sample configuration of the OFDM-related processors
5 1201A and 1201B and onward from Fig. 12. Components 1301A through 1310A
belong between 1201A and 312A from Fig. 12, while components 1301B through
1310B belong between 1201B and 3 12B.
[0 1661
Serial-to-parallel converter 1302A performs serial-to-parallel conversion on
10 weighted signal 1301A (corresponding to weighted signal 309A from Fig. 12) and
outputs parallel signal 1303A.
[0 1671
Reorderer 1304A takes parallel signal 1303A as input, performs reordering
thereof, and outputs reordered signal 1305A. Reordering is described in detail
15 later.
IFFT (Inverse Fast Fourier Transform) unit 1306A takes reordered signal
1305A as input, applies an IFFT thereto, and outputs post-IFFT signal 1307A.
[0168]
Wireless unit 1308A takes post-IFFT signal 1307A as input, performs
20 processing such as frequency conversion and amplification, thereon, and outputs
modulated signal 1309A. Modulated signal 1309A is then output as radio waves
by antenna 13 10A.
[O 1691
Serial-to-parallel converter 1302B performs serial-to-parallel conversion on
25 weighted signal 1301B (corresponding to post-phase change 309B fiom Fig. 12) and
outputs parallel signal 1303B.
[0 1701
Reorderer 1304B takes parallel signal 1303B as input, performs reordering
thereof, and outputs reordered signal 1305B. Reordering is described in detail
later.
IFFT unit 1306B takes reordered signal 1305B as input, applies an IFFT
5 thereto, and outputs post-IFFT signal 1307B.
[0171]
Wireless unit 1308B takes post-IFFT signal 1307B as input, performs
processing such as frequency conversion and amplification thereon, and outputs
modulated signal 1309B. Modulated signal 1309B is then output as radio waves by
10 antenna 13 10A.
[0 1 721
The transmission device from Fig. 3 does not use a multi-carrier
transmission scheme. Thus, as shown in Fig. 6, the change of phase is performed
to achieve a period (cycle) of four and the post-phase change symbols are arranged
15 with respect to the time domain. As shown in Fig. 12, when multi-carrier
transmission, such as OFDM, is used, then, naturally, precoded post-phase change
symbols may be arranged with respect to the time domain as in Fig. 3, and this
applies to each (sub-)carrier. However, for multi-carrier transmission, the
arrangement may also be in the frequency domain, or in both the frequency domain
20 and the time domain. The following describes these arrangements.
[O 1731
Figs. 14A and 14B indicate frequency on the horizontal axes and time on
the vertical axes thereof, and illustrate an example of a symbol reordering scheme
used by the reorderers 1301A and 1301B from Fig. 13. The fkequency axes are
25 made up of (sub-)carriers 0 through 9. The modulated signals zl and 22 share
common times (timing) and use a common frequency band. Fig. 14A illustrates a
reordering scheme for the symbols of modulated signal zl, while Fig. 14B illustrates
a reordering scheme for the symbols of modulated signal 22. With respect to the
49
symbols of weighted signal 1301A input to serial-to-parallel converter 1302A, the
assigned ordering is #0, #1, #2, #3, and so on. Here, given that the example deals
with a period (cycle) of four, #0, # 1, #2, and #3 are equivalent to one period (cycle).
Similarly, #4n, #4n+l, #4n+2, and #4n+3 (n being a non-zero positive integer) are
5 also equivalent to one period (cycle).
[0 1 741
As shown in Fig. 14A, symbols #0, #1, #2, #3, and so on are arranged in
order, beginning at carrier 0. Symbols #O through #9 are given time $1, followed
by symbols #10 through #19 which are given time #2, and so on in a regular
10 arrangement. Note that the modulated signals zl and 22 are complex signals.
[0 1 751
Similarly, with respect to the symbols of weighted signal 1301B input to
serial-to-parallel converter 1302B, the assigned ordering is #0, #1, #2, #3, and so on.
Here, given that the example deals with a period (cycle) of four, a different change
15 of phase is applied to each of #0, #1, #2, and #3, which are equivalent to one period
(cycle). Similarly, a different change of phase is applied to each of #4n, #4n+l,
#4n+2, and #4n+3 (n being a non-zero positive integer), which are also equivalent to
one period (cycle)
[0176]
20 As shown in Fig. 14B, symbols #O, #1, #2, #3, and so on are arranged in
order, beginning at carrier 0. Symbols #O through #9 are given time $1, followed
by symbols #10 through #19 which are given time #2, ahd so on in a regular
arrangement.
[0 1771
25 The symbol group 1402 shown in Fig. 14B corresponds to one period
(cycle) of symbols when the phase changing scheme of Fig. 6 is used. Symbol #O
is the symbol obtained by using the phase at time u in Fig. 6, symbol #I is the
symbol obtained by using the phase at time u+l in Fig. 6, symbol #2 is the symbol
50
obtained by using the phase at time u+2 in Fig. 6, and symbol #3 is the symbol
obtained by using the phase at time u+3 in Fig. 6. Accordingly, for any symbol #x,
symbol #x is the symbol obtained by using the phase at time u in Fig. 6 when x mod
4 equals 0 (i.e., when the remainder of x divided by 4 is 0, mod being the modulo
5 operator), symbol #x is the symbol obtained by using the phase at time u+l in Fig. 6
when x mod 4 equals 1, symbol #x is the symbol obtained by using the phase at time
u+2 in Fig. 6 when x mod 4 equals 2, and symbol #x is the symbol obtained by using
the phase at time u+3 in Fig. 6 when x mod 4 equals 3.
[0178]
10 In the present Embodiment, modulated signal zl shown in Fig. 14A has not
undergone a change of phase.
As such, when using a multi-carrier transmission scheme such as OFDM,
and unlike single carrier transmission, symbols may be arranged with respect to the
frequency domain. Of course, the symbol arrangement scheme is not limited to
15 those illustrated by Figs. 14A and 14B. Further examples are shown in Figs. 15A,
15B, 16A, and 16B.
[0 1791
Figs. 15A and 15B indicate frequency on the horizontal axes and time on
the vertical axes thereof, and illustrate an example of a symbol reordering scheme
20 used by the reorderers 1301A and 1301B from Fig. 13 that differs fiom that of Figs.
14A and 14B. Fig. 15A illustrates a reordering scheme for the symbols of
modulated signal zl, while Fig. 15B illustrates a reordering scheme for the symbols
of modulated signal 22. Figs. 15A and 15B differ fiom Figs. 14A and 14B in that
different reordering schemes are applied to the symbols of modulated signal zl and
25 to the symbols of modulated signal 22. In Fig. 15B, symbols #O through #5 are
arranged at carriers 4 through 9, symbols #6 though #9 are arranged at carriers 0
through 3, and this arrangement is repeated for symbols #10 through #19. Here, as
*
in Fig. 14B, symbol group 1502 shown in Fig. 15B corresponds to one period
(cycle) of symbols when the phase changing scheme of Fig. 6 is used.
[0 1 801
Figs. 16A and 16B indicate frequency on the horizontal axes and time on
5 the vertical axes thereof, and illustrate an example of a symbol reordering scheme
used by the reorderers 1301A and 1301B from Fig. 13 that differs from that of Figs.
14A and 14B. Fig. 16A illustrates a reordering scheme for the symbols of
modulated signal zl, while Fig. 16B illustrates a reordering scheme for the symbols
of modulated signal 22. Figs. 16A and 16B differ from Figs. 14A and 14B in that,
10 while Figs. 14A and 14B showed symbols arranged at sequential carriers, Figs. 16A
and 16B do not arrange the symbols at sequential carriers. Obviously, for Figs.
16A and 16B, different reordering schemes may be applied to the symbols of
modulated signal zl and to the symbols of modulated signal 22 as in Figs. 15A and
15B.
15 [0181]
Figs. 17A and 17B indicate frequency on the horizontal axes and time on
the vertical axes thereof, and illustrate an example of a symbol reordering scheme
used by the reorderers 1301A and 1301B from Fig. 13 that differs &om those of Figs.
14A through 16B. Fig. 17A illustrates a reordering scheme for the symbols of
20 modulated signal zl and Fig. 17B illustrates a reordering scheme for the symbols of
modulated signal 22. While Figs. 14A through 16B show symbols arranged with
respect to the frequency axis, Figs. 17A and 17B use the frequency and time axes
together in a single arrangement.
[0 1 821
25 While Fig. 6 describes an example where a change of phase is performed in
a four slot period (cycle), the following example describes an eight slot period
(cycle). In Figs. 17A and 17B, the symbol group 1702 is equivalent to one period
(cycle) of symbols when the phase changing scheme is used (i.e., to eight symbols)
52
such that symbol #O is the symbol obtained by using the phase at time u, symbol #1
is the symbol obtained by using the phase at time u+l, symbol #2 is the symbol
obtained by using the phase at time u+2, symbol #3 is the symbol obtained by using
the phase at time u+3, symbol #4 is the symbol obtained by using the phase at time
5 u+4, symbol #5 is the symbol obtained by using the phase at time u+5, symbol #6 is
the symbol obtained by using the phase at time u+6, and symbol #7 is the symbol
obtained by using the phase at time u+7. Accordingly, for any symbol #x, symbol
#x is the symbol obtained by using the phase at time u when x mod 8 equals 0,
symbol #x is the symbol obtained by using the phase at time u+l when x mod 8
10 equals 1, symbol #x is the symbol obtained by using the phase at time u+2 when x
mod 8 equals 2, symbol #x is the symbol obtained by using the phase at time u+3
when x mod 8 equals 3, symbol #x is the symbol obtained by using the phase at time
u+4 when x mod 8 equals 4, symbol #x is the symbol obtained by using the phase at
time u+5 when x mod 8 equals 5, symbol #x is the symbol obtained by using the
15 phase at time u+6 when x mod 8 equals 6, and symbol #x is the symbol obtained by
using the phase at time u+7 when x mod 8 equals 7. In Figs. 17A and 17B four
slots along the time axis and two slots along the frequency axis are used for a total
of 4x2 = 8 slots, in which one period (cycle) of symbols is arranged. Here, given
mxn symbols per period (cycle) (i.e., mxn different phases are available for
20 multiplication), then n slots (carriers) in the fiequency domain and m slots in the
time domain should be used to arrange the symbols of each period (cycle), such that
m > n. This is because the phase of direct waves fluctuates slowly in the time
domain relative to the frequency domain. Accordingly, the present Embodiment
performs a regular change of phase that reduces the influence of steady direct waves.
25 Thus, the phase changing period (cycle) should preferably reduce direct wave
fluctuations. Accordingly, m should be greater than n. Taking the above into
consideration, using the time and frequency domains together for reordering, as
shown in Figs. 17A and 17B, is preferable to using either of the frequency domain
5 3
or the time domain alone due to the strong probability of the direct waves becoming
regular. As a result, the effects of the present invention are more easily obtained.
However, reordering in the frequency domain may lead to diversity gain due the fact
that frequency-domain fluctuations are abrupt. As such, using the frequency and
5 time domains together for reordering is not always ideal.
[0183]
Figs. 18A and 18B indicate frequency on the horizontal axes and time on
the vertical axes thereof, and illustrate an example of a symbol reordering scheme
used by the reorderers 1301A and 1301B from Fig. 13 that differs from that of Figs.
10 17A and 14B. Fig. 18A illustrates a reordering scheme for the symbols of
modulated signal zl, while Fig. 18B illustrates a reordering scheme for the symbols
of modulated signal 22. Much like Figs. 17A and 17B, Figs. 18A and 18B
illustrate the use of the time and frequency domains, together. However, in
contrast to Figs. 17A and 17B, where the frequency domain is prioritized and the
15 time domain is used for secondary symbol arrangement, Figs. 18A and 18B
prioritize the time domain and use the frequency domain for secondary symbol
arrangement. In Fig. 18B, symbol group 1802 corresponds to one period (cycle) of
symbols when the phase changing scheme is used.
[0 1 841
20 In Figs. 17A, 17B, 18A, and 18B, the reordering scheme applied to the
symbols of modulated signal zl and the symbols of modulated signal 22 may be
identical or may differ as in Figs. 15A and 15B. Both approaches allow good
reception quality to be obtained. Also, in Figs. 17A, 17B, 18A, and 18B, the
symbols may be arranged non-sequentially as in Figs. 16A and 16B. Both
25 approaches allows good reception quality to be obtained.
[0185]
Fig. 22 indicates frequency on the horizontal axis and time on the vertical
axis thereof, and illustrates an example of a symbol reordering scheme used by the
54
reorderers 1301A and 1301B from Fig. 13 that differs fiom the above. Fig. 22
illustrates a regular phase changing scheme using four slots, similar to times u
through u+3 fiom Fig. 6. The characteristic feature of Fig. 22 is that, although the
symbols are reordered with respect the frequency domain, when read along the time
5 axis, a periodic shift of n (n = 1 in the example of Fig. 22) symbols is apparent.
The frequency-domain symbol group 2210 in Fig. 22 indicates four symbols to
which the change of phase is applied at times u through u+3 fiom Fig. 6.
[0186]
Here, symbol #O is obtained through a change of phase at time u, symbol #1
10 is obtained through a change of phase at time u+l, symbol #2 is obtained through a
change of phase at time u+2, and symbol #3 is obtained through a change of phase at
time u+3.
[0187]
Similarly, for fiequency-domain symbol group 2220, symbol #4 is obtained
15 through a change of phase at time u, symbol #5 is obtained through a change of
phase at time u+l, symbol #6 is obtained through a change of phase at time u+2, and
symbol #7 is obtained through a change of phase at time u+3.
[0188]
The above-described change of phase is applied to the symbol at time $1.
20 However, in order to apply periodic shifting in the time domain, the following phase
changes are applied to symbol groups 2201,2202,2203, and 2204.
[0189]
For time-domain symbol group 2201, symbol #O is obtained through a
change of phase at time u, symbol #9 is obtained through a change of phase at time
25 u+l, symbol # 18 is obtained through a change of phase at time u+2, and symbol #27
is obtained through a change of phase at time u+3.
[0 1901
For time-domain symbol group 2202, symbol #28 is obtained through a
change of phase at time u, symbol #1 is obtained through a change of phase at time
u+l, symbol #10 is obtained through a change of phase at time u+2, and symbol #19
is obtained through a change of phase at time u+3.
5 [0191]
For time-domain symbol group 2203, symbol #20 is obtained through a
change of phase at time u, symbol #29 is obtained through a change of phase at time
u+l, symbol #2 is obtained through a change of phase at time u+2, and symbol #11
is obtained through a change of phase at time u+3.
10 [0192]
For time-domain symbol group 2204, symbol #12 is obtained through a
change of phase at time u, symbol #21 is obtained through a change of phase at time
u+l, symbol #30 is obtained through a change of phase at time u+2, and symbol #3
is obtained through a change of phase at time u+3.
15 [0193]
The characteristic feature of Fig. 22 is seen in that, taking symbol #11 as an
example, the two neighbouring symbols thereof having the same time in the
frequency domain (#lo and #12) are both symbols changed using a different phase
than symbol # 1 1, and the two neighbouring symbols thereof having the same carrier
20 in the time domain (#2 and #20) are both symbols changed using a different phase
than symbol #I 1. This holds not only for symbol #I 1, but also for any symbol
having two neighboring symbols in the frequency domain and the time domain.
Accordingly, phase changing is effectively carried out. This is highly likely to
improve date reception quality as influence from regularizing direct waves is less
25 prone to reception.
[O 1 941
Although Fig. 22 illustrates an example in which n = 1, the invention is not
limited in this manner. The same may be applied to a case in which n = 3.
56
*
I Furthermore, although Fig. 22 illustrates the realization of the above-described
effects by arranging the symbols in the frequency domain and advancing in the time
domain so as to achieve the characteristic effect of imparting a periodic shift to the
symbol anangement order, the symbols may also be randomly (or regularly)
5 arranged to the same effect.
[0195]
[Embodiment 21
In Embodiment 1, described above, phase changing is applied to a weighted
(precoded with a fixed precoding matrix) signal z(t). The following Embodiments
10 describe various phase changing schemes by which the effects of Embodiment 1
may be obtained.
[0 1961
In the above-described Embodiment, as shown in Figs. 3 and 6, phase
changer 3 17B is configured to perform a change of phase on only one of the signals
15 output by the weighting unit 600.
However, phase changing may also be applied before precoding is
performed by the weighting unit 600. In addition to the components illustrated in
Fig. 6, the transmission device may also feature the weighting unit 600 before the
phase changer 3 17B, as shown in Fig. 25.
20 [0197]
In such circumstances, the following configuration is possible. The phase
changer 317B performs a regular change of phase with respect to baseband signal
s2(t), on which mapping has been performed according to a selected modulation
scheme, and outputs s2'(t) = s2(t)y(t) (where y(t) varies over time t). The
25 weighting unit 600 executes precoding on s2't, outputs z2(t) = W2s2'(t) (see Math.
42 (formula 42)) and the result is then transmitted.
[0198]
Alternatively, phase changing may be performed on both modulated signals
sl(t) and s2(t). As such, the transmission device is configured so as to include a
phase changer taking both signals output by the weighting unit 600, as shown in Fig.
26.
5 [0199]
Like phase changer 3 17B, phase changer 3 17A performs regular a regular
change of phase on the signal input thereto, and as such changes the phase of signal
zl'(t) precoded by the weighting unit. Post-phase change signal zl(t) is then output
to a transmitter.
10 [0200]
However, the phase changing rate applied by the phase changers 3 17A and
3 17B varies simultaneously in order to perform the phase changing shown in Fig. 26.
(The following describes a non-limiting example of the phase changing scheme.)
For time u, phase changer 3 17A from Fig. 26 performs the change of phase such that
15 zl (t) = yl(t)zl '(t), while phase changer 3 17B performs the change of phase such that
z2(t) = y2(t)z2'(t). For example, as shown in Fig. 26, for time u, yl(u) = do and
y2(u) = eYa, for time u+l, yl(u+l) = dnI4 and y2(u+l) = e - ~a~nd ~fo~r ti,m e u+k,
yl(u+k) = dknI4an d y2(u+k) = d(k3n14-a).H ere, the regular phase changing period
(cycle) may be the same for both phase changers 317A and 317l3, or may vary for
20 each.
[020 11
Also, as described above, a change of phase may be performed before
precoding is performed by the weighting unit. In such a case, the transmission
device should be configured as illustrated in Fig. 27.
25 [0202]
When a change of phase is carried out on both modulated signals, each of
the transmit signals is, for example, control information that includes information
about the phase changing pattern. By obtaining the control information, the
5 8
reception device knows the phase changing scheme by which the transmission
device regularly varies the change, i.e., the phase changing pattern, and is thus able
to demodulate (decode) the signals correctly.
[0203]
5 Next, variants of the sample configurations shown in Figs. 6 and 25 are
described with reference to Figs. 28 and 29. Fig. 28 differs fkom Fig. 6 in the
inclusion of phase change ONIOFF information 2800 and in that the change of
phase is performed on only one of zlr(t) and z2'(t) (i.e., performed on one of zl'(t)
and z2'(t), which have identical times or a common frequency). Accordingly, in
10 order to perform the change of phase on one of zl '(t) and z2'(t), the phase changers
317A and 317B shown in Fig. 28 may each be ON, and performing the change of
phase, or OFF, and not performing the change of phase. The phase change
ONIOFF information 2800 is control information therefor. The phase change
ONIOFF information 2800 is output by the signal processing scheme information
15 generator 3 14 shown in Fig. 3.
[0204]
Phase changer 317A of Fig. 28 changes the phase to produce zl(t) =
yl(t)zlf(t), while phase changer 317B changes the phase to produce z2(t) =
~2(t)z2'(t).
20 [0205]
Here, a change of phase having a period (cycle) of four is, for example,
applied to zl'(t). (Meanwhile, the phase of z2'(t) is not changed.) Accordingly, for
time u, y,(u) = e'' and y2(u) = 1, for time u+l, yl(u+l) = Pn and y2(u+l) = 1, for
time u+2, yl(u+2) = e'" and y2(u+2) = 1, and for time u+3, yl(u+3) = e'3n'2 and
25 y2(u+3) = 1.
[0206]
Next, a change of phase having a period (cycle) of four is, for example,
applied to z2'(t). (Meanwhile, the phase of zll(t) is not changed.) Accordingly, for
5 9
time u+4, yl(u+4) = I and y2(u+4) = 2, for time u+5, yl(u+5) = 1 and y2(u+5) = dd2,
for time u+6, yl(u+6) = 1 and y2(u+6) = c, and for time u+7, yl(u+7) = 1 and
y2(u+7) = e'3"n.
[0207]
5 Accordingly, given the above examples.
for any time 8k, y1(8k) = 8 and y2(8k) = 1,
for any time 8k+l, y1(8k+l) = ddandy2(8k+l) = 1,
for any time 8k+2, y1(8k+2) = e'" and y2(8k+2) = 1,
for any time 8k+3, y1(8k+3) = d3d and y2(8k+3) = 1,
10 for any time 8k+4, y1(8k+4) = 1 and y2(8k+4) = e",
for any time 8k+5, yl(8k+3) = 1 and y2(8k+5) = dd,
for any time 8k+6, y1(8k+6) = 1 and y2(8k+6) = e'", and
for any time 8k+7, y1(8k+7) = 1 and y2(8k+7) = e'3"/2.
[0208]
15 As described above, there are two intervals, one where the change of phase
is performed on zll(t) only, and one where the change of phase is performed on
z2'(t) only. Furthermore, the two intervals form a phase changing period (cycle).
While the above explanation describes the interval where the change of phase is
performed on zll(t) only and the interval where the change of phase is performed on
20 z2'(t) only as being equal, no limitation is intended in this manner. The two
intervals may also differ. In addition, while the above explanation describes
performing a change of phase having a period (cycle) of four on zl'(t) only and then
performing a change of phase having a period (cycle) of four on z2'(t) only, no
limitation is intended in this manner. The changes of phase may be performed on
25 zl'(t) and on z2'(t) in any order (e-g., the change of phase may alternate between
being performed on zll(t) and on z2'(t), or may be performed in random order).

CLAIMS
1. A signal generation method for generating, from a plurality of baseband signals, a
plurality of signals for transmission on a common frequency band and at a common
time, comprising the steps of:
5 generating M first encoded blocks usable as a first set of bits and M second
encoded blocks usable as a second set of bits using a predetermined error-correcting
block coding method, where M is a natural number;
performing a change of phase on each of a first baseband signal si
generated from the first set of bits and a second baseband signal s2 generated from
10 the second set of bits, thus generating a first post-phase change baseband signal si'
and a second post-phase change baseband signal s2', each including M symbols; and
applying weighting to the first post-phase change baseband signal si' and to
the second post-phase change baseband signal s2' according to a predetermined
matrix F, thus generating the plurality of signals for transmission on the common
15 frequency band and at the common time as a combination of M pairs of a first
weighted signal zl and a second weighted signal z2, wherein
the first weighted signal zl and the second weighted signal z2 satisfy the
relation:
(zl,z2f = F(sl',s2'f
20 and the change of phase is performed on the first baseband signal si and the
second baseband signal s2 using a phase modification value sequentially selected
from among N phase modification value candidates.
2. A signal generation apparatus for generating, from a plurality of baseband signals,
25 a plurality of signals for transmission on a common frequency band zind at a
common time, comprising:
an encoder generating M first encoded blocks usable as a first set of bits and
M second encoded blocks usable as a second set of bits using a predetermined
296
•
error-correcting block coding method, where M is a natural number;
a phase changer performing a change of phase on each of a first bziseband
signal si generated from the first set of bits and a second baseband signal s2
generated from the second set of bits, thus generating a first post-phase change
5 baseband signal si' and a second post-phase change baseband signal s2', each
including M symbols; and
a weighting unit applying weighting to the first post-phase change baseband
signal si' and to the second post-phase change baseband signal s2' according to a
predetermined matrix F, thus generating the plurality of signals for transmission on
10 the common frequency band and at the common time as a combination of M pairs of
a first weighted signal zl and a second weighted signal z2, wherein
the first weighted signal zl and the second weighted signal z2 satisfy the
relation:
(zl, z2f = F(sl', s 2 '/
15 and the change of phase is performed on the first baseband signal si and the
second baseband signal s2 using a phase modification value sequentially selected
from among N phase modification value candidates.