JP2005039585A - Information transmission method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an information transmission method which allows information encoded by using irregular LDPC codes to be so transmitted that BER characteristics at the time of decoding are improved. <P>SOLUTION: When encoded information obtained by encoding information on a transmission object by using a generator matrix generated from an irregular LDPC matrix is transmitted, the phase of a signal to be outputted (transmission signal) is switched (BPSK modulation) in accordance with the value of each bit of the encoded information, and the amplitude of a carrier (transmission signal) is changed in accordance with weights of corresponding columns of the irregular LDPC matrix (refer to (B)) so as to increase/reduce transmission powers assigned to bits corresponding to columns of light/heavy weights of the irregular LDPC matrix, out of individual bits of the encoded information when modulating the signal in accordance with values of individual bits. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information transmission method and apparatus, and in particular, an information transmission method for encoding and transmitting information to be transmitted using a low density parity check code defined by a low density parity check matrix, and the information transmission method. The present invention relates to an applicable information transmission apparatus.

In recent years, a low density parity check (LDPC) code has been attracting attention as an excellent error correction code exhibiting characteristics close to the Shannon limit as in the case of a turbo code. The LDPC code is a linear block code defined by a very sparse parity check matrix (low density parity check matrix: LDPC matrix). A sparse matrix refers to a matrix with a very large number of zeros in the matrix and a very small number of ones. An example of an LDPC code (LDPC matrix H ₁ ) is shown below.

The decoding characteristics of LDPC codes are comparable to turbo codes, or when the code length is sufficiently long, a characteristic result has been reported that the error floor phenomenon observed in the decoding characteristics of turbo codes etc. (The phenomenon in which the bit error rate hardly changes even if the error occurs) has an advantage that it hardly occurs.

Also, the LDPC code can be expressed using a Factor graph. The Factor graph represents an LDPC matrix of M rows and N columns by M check nodes, N bit nodes, and edges connecting them. As an example, FIG. 14 shows a factor graph of the LDPC matrix H ₁ described above. Since the LDPC matrix H ₁ has 4 rows and 6 columns, the Factor graph in FIG. 14 is composed of 4 check nodes and 6 bit nodes. Each bit node of the Factor graph corresponds to each bit (or each column of the LDPC matrix) of encoded information (codeword) obtained by encoding information using the LDPC matrix H ₁ , and each check node Corresponds to each row of the LDPC matrix. Further, the edge existing between a certain bit node and a certain check node has 1 at the intersection position between the column of the LDPC matrix corresponding to the bit node and the row of the LDPC matrix corresponding to the check node. Means.

  As a characteristic of the LDPC code, the code is configured to be “pseudo-random”. As guaranteed by the channel coding theorem, when the code length is sufficiently long, the code selected from the random code ensemble almost certainly has excellent maximum likelihood decoding performance. However, since the random code does not have a structure that assists efficient decoding, the time required for decoding increases exponentially as the code length increases. On the other hand, since the LDPC code has a low density (sparse) check matrix, the LDPC code has a “sparse graph structure” that is a structure suitable for decoding while maintaining randomness that provides excellent maximum likelihood decoding characteristics.

Several proposals have been made regarding LDPC codes. For example, Patent Document 1 assigns the same number of non-zero symbols to all rows of a check matrix A when constructing an LDPC code, and the number of symbols t. Is defined as such that any two columns of the check matrix A have at most one non-zero value, and only two rows of the check matrix A are present. It has been proposed to define rows in such a way as to have a non-zero common value.
JP 2001-168733 A

  By the way, the LDPC code includes a regular LDPC code and an irregular LDPC code. The regular LDPC code is an LDPC code defined by a check matrix in which the weight of each row (the number of 1s in each row) and the weight of each column (the number of 1s in each column) are fixed. In the Factor graph, the number of edges connected to each bit node and the number of edges connected to each check node are all equal.

  On the other hand, an irregular LDPC code is an LDPC code defined by an LDPC matrix in which the weight of each row and the weight of each column are not constant. In the Factor graph, the number of edges connected to each bit node and the number of edges connected to each check node are different. It is known that if an irregular LDPC code having an appropriate distribution of weights in each row and weight in each column is used, a better BER (Bit Error Rate) characteristic can be obtained than when a regular LDPC code is used. ing.

  On the other hand, in decoding of information encoded by the LDPC code, a BP (Belief Propagation) algorithm, which is a typical algorithm of the iterative decoding method, is used in advance for all rows (check nodes) other than the bit of interest. After calculating the probability that the parity check sum is satisfied from the probability, the probability that all the parity check sums related to the bit of interest are satisfied as the likelihood of the bit of interest (the parity check sum of the row having 1 in the column of interest) (Probability that is 0) is calculated for all columns (bit nodes), and the algorithm is repeated to replace the likelihood calculated for each bit with a prior probability. By repeatedly updating the likelihood as described above, a more accurate likelihood of each bit is calculated, and the encoded information is decoded.

  As described above, in the irregular LDPC code, the weight of each column (and each row) is not constant. For this reason, among the encoded information encoded using the irregular LDPC code, the bit corresponding to the column having a small weight in the LDPC matrix (the number of edges connected to the corresponding bit node in the above-described Factor graph is small) Bits), the number of pieces of information that can be used for calculating the likelihood at the time of decoding is smaller than the bits corresponding to the columns having a large weight in the LDPC matrix, so that the reliability of error correction is low and the BER characteristics are improved. It was a cause to inhibit.

  The present invention has been made in consideration of the above facts, and is an information transmission method and information transmission capable of transmitting information encoded using an irregular LDPC code so that the BER characteristics at the time of decoding are improved. The purpose is to obtain a device.

  In order to achieve the above object, the information transmission method according to the first aspect of the present invention provides information to be transmitted using a low density parity check code defined by a low density parity check matrix in which the weight of each row and each column is not constant. Of the encoded information obtained by encoding the transmission target information, the bit corresponding to the column having a small weight in the low density parity check matrix is the low density. The encoded information is transmitted so that the probability of receiving without error on the receiving side is higher than that of a bit corresponding to a column having a large weight in the parity check matrix.

In the first aspect of the invention, information to be transmitted is transmitted using a low density parity check code (regular LDPC code) defined by a low density parity check matrix (irregular LDPC matrix) in which the weight of each row and each column is not constant. Is encoded and transmitted. Incidentally, encoding of information using irregular LDPC code, for example, irregular LDPC matrix H and the following (1) to determine the product matrix G having a relation of equation (where, H ^T is a transposed matrix of the matrix H),
GH ^T = 0 (1)
It can be realized by encoding the information u to be transmitted by the generated generator matrix G (see the following equation (2)) (s: encoded information (transmission word vector)).
s = uG (2)
As described above, the information encoded using the irregular LDPC code has a problem that the reliability of error correction for the bit corresponding to the column having a small weight in the irregular LDPC matrix is low at the time of decoding. Therefore, according to the first aspect of the present invention, among the bits of the encoded information obtained by encoding the information to be transmitted, the bit corresponding to the column having a small weight in the irregular LDPC matrix is the irregular LDPC. The encoded information is transmitted so that the probability of receiving without error on the receiving side is higher than that of the bit corresponding to the column having a large weight in the matrix.

  Thus, on the receiving side that receives the encoded information transmitted by applying the information transmitting method according to the first aspect of the invention, the received encoded information corresponds to a column having a small weight in the irregular LDPC matrix. By improving the probability that a bit is received without error, low reliability of error correction for the bit can be compensated. Therefore, according to the first aspect of the present invention, it is possible to transmit information encoded using the irregular LDPC code so that the BER characteristic at the time of decoding is improved.

  In the first aspect of the present invention, among the bits of the encoded information, the bit corresponding to the column having a small weight in the irregular LDPC matrix is more than the bit corresponding to the column having a large weight in the irregular LDPC matrix. Specifically, the transmission of the encoded information so that the probability of being received without error on the receiving side is high, specifically corresponds to a column with a small weight in the irregular LDPC matrix as described in claim 2, for example. This can be realized by increasing the transmission power allocated to the bits to be performed. Thus, on the receiving side that receives the encoded information, the encoded information is transmitted so that the probability that the bit corresponding to the column having a small weight in the irregular LDPC matrix among the encoded information is received without error is improved. Can do.

  Further, in the invention described in claim 2, among the bits of the encoded information, for the bit corresponding to the column having a large weight in the irregular LDPC matrix, the transmission power to be allocated is reduced as described in claim 3, for example. It is preferable to make it. Thereby, by applying this invention, it can suppress that the electric power required in order to transmit encoding information increases.

  Further, in the invention according to claim 2 or claim 3, the transmission power is assigned to each bit of the encoded information by, for example, determining whether the weight of the corresponding column of the irregular LDPC matrix is equal to or greater than a threshold value. May be determined by switching the transmission power allocated to each bit in two stages according to the result of the determination. For example, as described in claim 4, Preferably, the transmission power assigned to each bit of information is made inversely proportional to the square of the weight of the corresponding column in the irregular LDPC matrix. Thereby, the ratio of the transmission power allocated to each bit of the encoded information can be optimized according to the weight of the corresponding column of the irregular LDPC matrix.

  Furthermore, in the invention of claim 2 or claim 3, the transmission power assigned to each bit of the encoded information is changed (the bit corresponding to the column having a small weight in the irregular LDPC matrix among the bits of the encoded information). (E.g., increasing the allocated transmit power or increasing the transmit power allocated to the bit corresponding to the low weight column and decreasing the transmit power allocated to the bit corresponding to the large weight column). As described above, it can be realized by partially changing the amplitude of the carrier wave modulated in accordance with the encoded information.

  Further, in the invention according to claim 2 or claim 3, changing the transmission power allocated to each bit of the encoded information modulates the carrier according to the encoded information as described in claim 6, for example. This can also be realized by transmitting a portion modulated in accordance with the bit corresponding to the column having a small weight among the transmission signals obtained in this way.

  In the first aspect of the present invention, among the bits of the encoded information, the bit corresponding to the column having a small weight in the irregular LDPC matrix is more than the bit corresponding to the column having a large weight in the irregular LDPC matrix. For example, as described in claim 7, the transmission of the encoded information so as to increase the probability of receiving without error on the receiving side is performed by changing the bit corresponding to the column having a small weight in the irregular LDPC matrix to the frequency. It transmits using a channel having a good communication state among a plurality of channels separated by at least one of a region, a time region, a space region, and a code region, and corresponds to a column having a large weight in the irregular LDPC matrix A communication path different from the transmission path used for transmitting the bit corresponding to the column having a small weight in the irregular LDPC matrix It can also be achieved by transmitting using. Also in this case, on the receiving side that receives the encoded information, the encoded information can be transmitted so that the probability that the bit corresponding to the column having a small weight in the irregular LDPC matrix among the encoded information is received without error is improved. .

  In the invention according to claim 7, when the communication states of each of the plurality of communication paths are unknown, for example, as described in claim 8, each of the known signals is transmitted using the plurality of communication paths. The communication state of each of the plurality of communication paths can be detected based on the reception state on the reception side of the known signal transmitted via each of the plurality of communication paths.

  The information transmitting apparatus according to the ninth aspect of the invention uses a low density parity check code (irregular LDPC code) defined by a low density parity check matrix (irregular LDPC matrix) in which the weight of each row and each column is not constant. An information transmission apparatus that encodes and transmits information to be transmitted, and has a small weight in the low-density parity check matrix among the bits of the encoded information obtained by encoding the information to be transmitted Transmission control means for transmitting the coded information so that the probability that a bit corresponding to a column is received without error on the receiving side is higher than a bit corresponding to a column having a large weight in the low-density parity check matrix As in the first aspect of the present invention, the information encoded using the irregular LDPC code is decoded at the time of decoding. It can be transmitted as ER characteristics are improved.

  As described above, the present invention encodes and transmits information to be transmitted using an irregular LDPC code, and among the bits of the encoded information, bits corresponding to a column having a small weight in the LDPC matrix are: Since the encoded information is transmitted so that the probability of receiving without error on the receiving side is higher than the bit corresponding to the column having a large weight in the LDPC matrix, encoding is performed using the irregular LDPC code. It has an excellent effect that information can be transmitted so that the BER characteristic at the time of decoding is improved.

  Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 shows a transmission device 10 for transmitting information to be transmitted, and FIG. 2 shows a reception device 20 for receiving information transmitted from the transmission device 10. The transmitter 10 corresponds to the information transmitter according to the ninth aspect of the invention. In the following, an example in which communication between the transmission device 10 and the reception device 20 is performed wirelessly will be described, but it goes without saying that the present invention is also applicable to wired communication.

  The transmission apparatus 10 is configured by connecting an LDPC encoder 12, a modulator 14, and an antenna 16 in series. The LDPC encoder 12 includes a storage unit. The storage unit includes an M-row × N-column regular LDPC matrix H determined so that the weight (number of 1s) of each row and each column is not constant. A generation matrix G of K rows × N columns obtained in accordance with the above-described equation (1) based on the irregular LDPC matrix H is stored in advance. Transmission target information (transmission word) is sequentially input to the LDPC encoder 12 in units of K bits, and the LDPC encoder 12 uses the generation matrix G stored in the storage unit for the input transmission target information. The N-bit encoded information (code word) s is generated and output by performing the calculation of the above equation (2). The encoded information includes a transmission word and a check word as information.

  The modulator 14 modulates a carrier wave according to the N-bit encoded information input from the LDPC encoder 12 to generate a transmission signal for transmitting the encoded information wirelessly via the antenna 16. Also, the modulator 14 increases the transmission power assigned to the bit corresponding to the column having a small weight in the irregular LDPC matrix among the bits of the input encoded information as compared with the case where the uniform transmission power is assigned to each bit. In addition, the transmission power assigned to the bit corresponding to the column having a large weight in the irregular LDPC matrix is reduced as compared with the case where the uniform transmission power is assigned to each bit. The transmission signal output from the modulator 14 is supplied to the antenna 16 and radiated from the antenna 16 as an electromagnetic wave modulated according to the encoded information.

  On the other hand, the receiving apparatus 20 shown in FIG. 2 is configured by connecting an antenna 22, a demodulator 24, and an LDPC decoder 26 in series. When an electromagnetic wave radiated from the antenna 16 of the transmission device 10 is received by the antenna 22 and a reception signal is output from the antenna 22, the demodulator 24 encodes N-bit encoded information based on the input reception signal. (Received word) r is demodulated.

  By the way, when there is no noise in the communication path between the transmission device 10 and the reception device 20, the received word r = codeword s output from the demodulator 24 is obtained, but actually, from the antenna 16 of the transmission device 10. Due to the noise superimposed until the radiated electromagnetic wave is received by the antenna 22 of the receiver 20, the noise is also superimposed on the received word r output from the demodulator 24 (r = s + n, where n Is a noise component).

There is a relationship represented by the following equation (3) between the code word s and the irregular LDPC matrix H used for encoding the code word s.
sH ^T = 0 (3)
For this reason, the LDPC decoder 26 estimates the original codeword by performing decoding (error correction) of the received word r, and determines whether or not the expression (3) holds from the codeword estimation result s ′. Then, it is verified whether or not the original code word is obtained. Then, the transmission word (the estimation result) is extracted from the code word estimation result s ′ and output.

  Next, as an operation of the present embodiment, first, a modulation process performed by the modulator 14 of the transmission apparatus 10 will be described. In the following description, for simplicity of explanation, it is assumed that BPSK (Binary Phase Shift Keying) modulation, which is one of digital modulation methods using phase modulation, is performed as a modulation method in the modulator 14.

  As shown in FIG. 3A, in normal BPSK modulation, the phase of a carrier wave having a constant amplitude is assigned to 0 and the phase opposite to the carrier wave (phase shifted by 180 °) is assigned to 1, and information to be transmitted (this embodiment) In this embodiment, the carrier wave is modulated (the phase of the signal to be output is switched) in accordance with the value of each bit of the N-bit encoding information input from the LDPC encoder 12. On the other hand, in the transmission apparatus 10 according to the present embodiment, the LDPC encoder 12 generates and outputs N-bit encoded information, and at the same time, the correspondence of the irregular LDPC matrix H to each bit of the encoded information to be output. Column weight information indicating the column weight (number of 1s in the column) is also output.

  The modulator 14 switches the phase of the output signal (transmission signal) in accordance with the value of each bit of the encoded information input from the LDPC encoder 12 and, at the same time, the column weight input from the LDPC encoder 12. Based on the information, as shown in FIG. 3B as an example, when modulating each bit of the encoded information according to the bit corresponding to the column having a small weight in the irregular LDPC matrix H, a carrier wave (transmission signal) ) And the amplitude of the carrier wave (transmission signal) according to the size of the column weight so that the amplitude of the carrier wave (transmission signal) decreases when modulating according to the bit corresponding to the column with a large weight. Change.

  As a result, the transmission power allocated to the bit corresponding to the column having a small weight in the irregular LDPC matrix H is increased and the column having a large weight in the irregular LDPC matrix H is hardly changed without substantially changing the transmission power of the entire transmission signal. The transmission power allocated to the bits corresponding to can be reduced.

Note that changing the amplitude (transmission power) of the carrier wave according to the size of the column weight is specifically, for example, changing the amplitude (transmission power) of the carrier wave in two stages depending on whether the column weight is equal to or greater than a threshold value. However, the transmission power assigned to each bit of the encoded information is inversely proportional to the square of the corresponding column weight (number of 1s) of the irregular LDPC matrix H (ie, 1). It is preferable to change the amplitude (transmission power) of the carrier wave so as to be proportional to / w ² (w is a column weight). Thereby, the ratio of the transmission power allocated to each bit of the encoded information can be optimized according to the weight of the corresponding column of the irregular LDPC matrix H.

  Further, when the column weight of the irregular LDPC matrix H corresponding to each bit of the encoded information is each value, the optimum magnitude of the transmission power assigned to each bit is defined by the irregular LDPC matrix H. Since it varies depending on the configuration (column weight, row weight and size) of irregular LDPC code, communication channel, and modulation method, the column weight has each value according to the target communication method, communication channel, and LDPC code to be used. It is desirable to calculate and store in advance the amount of transmission power assigned to each bit, and to read out the calculation result during modulation and use it for transmission power control.

  The transmission signal output from the modulator 14 by the above-described modulation processing being performed by the modulator 14 is supplied to the antenna 16 and is radiated from the antenna 16 as an electromagnetic wave. The electromagnetic wave radiated from the antenna 16 is received by the antenna 22 of the receiving device 20, and the demodulator 24 demodulates and outputs N-bit encoded information (received word) r based on the received signal input from the antenna 22. Demodulation processing is performed. The encoded information (received word) r output from the demodulator 24 is input to the LDPC decoder 26.

  Next, the decoding process performed by the LDPC decoder 26 will be described. This decoding process is a process to which the BP algorithm, which is one of the LDPC code iterative decoding methods, is applied. First, the following notation will be described.

N (m) = [n; H _mn = 1] is a set of bit nodes connected to the check node m by an edge on the Factor graph, that is, the position of 1 in the mth row of the irregular LDPC matrix H. means. M (n) = [m; H _mn = 1] means a set of check nodes connected to the bit node n, that is, the position of 1 in the nth column of the irregular LDPC matrix H. N (m) \ n means N (m) excluding the nth column of the irregular LDPC matrix H, and M (n) \ m denotes M (excluding the mth row of the irregular LDPC matrix H. n). FIG. 4 shows specific examples of N (m), M (n), N (m) \ n, and M (n) \ m.

The BP algorithm is an algorithm that repeatedly updates and decodes the bit probability information q _mn and r _mn using two parts, a bit node and a check node. q ⁰ _mn and q ¹ _mn are probability information (see also FIG. 5) that the n-th bit node sends to the m-th check node connected by an edge on the Factor graph, and each P (x _n = 0). And P (x _n = 1). Similarly, r ⁰ _mn and r ¹ _mn represent probability information (see also FIG. 6) that the m-th check node sends to the n-th bit node connected by an edge on the Factor graph.

That is, r ⁰ _mn (r ¹ _mn ) uses the probability information q _{mn ′} other than the target bit x _n for the _target bit x _n = 0 (x _n = 1), so that the m-th parity check sum is obtained. It represents the probability of being satisfied (the probability of becoming 0). Therefore, r ⁰ _mn is considered as external information from the mth check node to the nth bit node. As shown in FIG. 6, the posterior probability for a certain bit is calculated by collecting external information from all check nodes connected to a certain bit (node) by an edge on the Factor graph. In the following, it is assumed that the probability of being 0 among the initial prior probabilities of bit x _n is f ⁰ _n and the probability of ¹ is f ¹ _n .

Hereinafter, the decoding process performed by the LDPC decoder 26 will be described with reference to the flowchart of FIG. In step 100, initial prior probabilities f ⁰ _n and f ¹ _n are assigned to all variables q ⁰ _mn and q ¹ _mn corresponding to all combinations of m and n (corresponding to individual edges on the Factor graph), respectively. Perform initialization. In the next step 102, the following calculation is performed as row processing. That is,
δq _mn = q ⁰ _mn −q ¹ _mn
The following equations (4) to (6) are calculated for all combinations of m and n.

r ⁰ _mn = (1/2) (1 + (− 1) ⁰ δr _mn ) (5)
r ¹ _mn = (1/2) (1 + (− 1) ¹ δr _mn ) (6)
By the above processing, for example, probability information q ^x _{m1n1 to} q ^x _m1n3 is propagated from the bit nodes n1 to n3 to the check node m1 connected to the bit nodes n1 to n3 at the edge on the Factor graph as shown in FIG. The probability information r ^x _m1n2 to be propagated from the check node m1 to the bit node n2 is calculated using the probability information q ^x _m1n1 and q ^x _m1n3 propagated from the bit nodes n1 and n3 ( _note that x Means bit value = 0,1).

In step 104, the following calculation is performed as column processing. That is, the posterior probabilities q ⁰ _mn and q ¹ _mn are updated according to the following equations (7) and (8) for all combinations of m and n.

Α _mn is selected so as to satisfy q ⁰ _mn + q ¹ _mn = 1. In the column processing, q ⁰ _n and q ¹ _n are updated for all _{n according} to the following equations (9) and (10).

Note that α _n is selected so as to satisfy q ⁰ _n + q ¹ _n = 1.

As a result of the above processing, the probability information r ^x _{m1n1 to} r ^x _m3n1 is propagated from the check nodes m1 to m3 to the bit node n1 connected to the check nodes m1 to m3 at the edge on the Factor graph, for example, as shown in FIG. The probability information q ^x _m2n2 to be propagated from the bit node n1 to the check node m2 is calculated using the probability information r ^x _m1n1 and r ^x _m3n1 propagated from the check nodes m1 and m3.

In step 106, end determination is performed. That is, if q ¹ _n > 0.5, x ′ _n = 1, and if q ¹ _n ≦ 0.5, x ′ _n = 0 is performed for all the bit nodes, so that the bit string x ′ = [x ′ _n ] is generated. The bit string x ′ corresponds to a codeword estimation result. Then, it is determined whether or not the following expression (11) holds for the generated bit string x ′.
Hx ′ = 0 (11)
If the expression (11) is satisfied, it can be determined that the bit string x ′ is an effective estimation result of the code word, so the determination in step 106 is affirmed, the bit string x ′ is stored, and the decoding process is terminated.

On the other hand, if the expression (11) is not established, it is determined whether or not the number of executions of step 102 to step 106 has reached a predetermined maximum number of repetitions. If this determination is negative, it can be determined that the decoding process has not been completed, so the process returns to step 102 and the processes after step 102 are repeated. As a result, the process of sequentially updating the probability information r ^x _mn , q ^x _mn , q ^x _n and determining whether the bit string x ′ generated from the updated probability information satisfies the expression (11) is repeated. become. When the expression (11) is satisfied or the number of executions of steps 102 to 106 reaches the maximum number of repetitions, the determination in step 106 is affirmed and the decoding process is terminated.

As described above, in the column process (step 104) of the decoding process, the probability information q ^x _mn , q ^x is based on the probability information r ^x _mn propagated from the check node connected to the bit node by the edge on the Factor graph. _{Since n} is calculated (updated), the bit node corresponding to the column having a small weight on the irregular LDPC matrix H has the probability information q ^x based on the probability information r ^x _mn propagated from a small number of check nodes. _{Since mn} and q ^x _n are calculated (updated), there is a problem that the reliability of the probability information q ^x _n used for generating the bit string x ′ is low.

  On the other hand, in the present embodiment, when transmitting the encoded information obtained by encoding the transmission target information using the generation matrix G generated from the irregular LDPC matrix H, among the bits of the encoded information, Since a large transmission power is assigned to a bit corresponding to a column with a small weight on the irregular LDPC matrix H, the noise superimposed while being transmitted through the communication path between the transmission device 10 and the reception device 20 In the encoded information (received word) output from the demodulator 24 of the receiving apparatus 20 with the influence reduced, there is a probability that the bit value corresponding to the column with a small weight on the irregular LDPC matrix H is changed. Reduced. Therefore, the BER characteristic at the time of decoding can be improved.

  In the above, BPSK modulation has been described as an example of the modulation method in the modulator 14, but the present invention is not limited to this, and it goes without saying that various known modulation methods can be applied. For example, when an OFDM (Orthogonal Frequency Division Multiplexing) modulation method is applied as a modulation method in the modulator 14, the modulator 14 receives input encoded information as shown in FIG. The mapping / power allocating unit 30 that performs conversion to complex symbol sequences and performs transmission power allocation as described above, and serial-parallel conversion that converts input complex symbol sequences into a plurality of low-speed (same number as the number of OFDM carriers) symbol sequences. Unit 32, an inverse discrete Fourier transform (IDFT) that performs inverse discrete Fourier transform (IDFT) on a plurality of input symbol sequences, and superimposes each symbol sequence on each carrier, and converts the input parallel signal into a serial signal. Parallel to serial conversion unit 36 that performs D / A (digital to analog) conversion, low pass filter (LPF) 38, input Signal (the complex baseband OFDM signal) can be configured by the frequency converter 40 and a band pass filter (BPF) 42 is converted into a signal in the RF band.

  In the above configuration, the transmission power allocated to the bit corresponding to the column having a large weight in the irregular LDPC matrix H is increased by increasing the transmission power allocated to the bit corresponding to the column having a small weight in the irregular LDPC matrix H. The reduction can be performed by the mapping / power allocation unit 30. Also, in the mapping / power allocation unit 30, when the input encoded information is converted (assigned) into a complex symbol sequence, an irregular LDPC matrix H is placed at a symbol position having a large degree of contribution to the peak power of the OFDM signal. A symbol mapped with a bit corresponding to a column having a large weight (a bit to which a small power is allocated) is allocated to a symbol position where the degree of contribution to the peak power of the OFDM signal is small, and to a column having a small weight in the irregular LDPC matrix H A symbol to which a corresponding bit (bit for allocating high power) is mapped is allocated. Thereby, in addition to the improvement of the BER characteristic at the time of decoding, suppression of the peak power of the OFDM signal can be realized at the same time.

  When the OFDM modulation method is applied as the modulation method in the modulator 14, the demodulator 24, for example, as shown in FIG. 10B, the BPF 46, the frequency at which the input RF band signal is converted into a complex baseband OFDM signal. The conversion unit 48, the LPF 50, A / D (analog-to-digital) conversion for the input signal and serial / parallel conversion unit 52 for generating parallel data for each symbol, discrete Fourier transform for the input parallel data A discrete Fourier transform unit 54 that performs transformation (DFT), a parallel / serial conversion unit 56 that converts a plurality of input symbol sequences into a serial complex symbol sequence, and a data converter that converts the input complex symbol sequence into encoded information. What is necessary is just to comprise by the mapping part 58. FIG.

  Further, in transmission apparatus 10 shown in FIG. 1, transmission power allocated to bits corresponding to a column with a small weight in irregular LDPC matrix H in the encoded information is increased, and transmission allocated to a bit corresponding to a column with a large weight is transmitted. The power reduction is realized by changing the amplitude of the carrier wave modulated in accordance with the encoded information. However, the present invention is not limited to this, and in the irregular LDPC matrix H of the encoded information. Increasing the transmission power assigned to the bit corresponding to the column having a small weight to the transmission power assigned to the bit corresponding to the column having a large weight corresponds to the column having a small weight in the irregular LDPC matrix H in the encoded information. When the symbols mapped with the bit to be transmitted are transmitted multiple times, and the symbols transmitted multiple times are added together Can also be achieved by transmitting power is greater than the other symbols.

  Transmitting a symbol mapped with bits corresponding to a column having a small weight in the irregular LDPC matrix H among the encoded information a plurality of times means that the symbol appears in the transmission signal a plurality of times or the transmission word vector s or the transmission This can be realized by processing the signal or retransmitting only the symbol after transmitting the transmission signal once. In this case, as shown in FIG. 11, a correlator 28 may be provided between the antenna 22 and the demodulator 24 of the receiving device 20, and the same symbol may be added by the correlator 28.

  Also, in the above, among the bits of the encoded information, the bit corresponding to the column with a small weight in the irregular LDPC matrix H is more erroneous on the receiving side than the bit corresponding to the column with a large weight in the irregular LDPC matrix H. Transmission of encoded information so as to increase the probability of being received without increasing the transmission power allocated to bits corresponding to a column having a small weight in the irregular LDPC matrix H and the weight of the irregular LDPC matrix H. Although the aspect realized by reducing the transmission power allocated to the bit corresponding to the large column has been described, the present invention is not limited to this. For example, as illustrated in FIG. A plurality (separated between at least one of a frequency domain, a time domain, a spatial domain, and a code domain between the devices 20 ( If there are M communication channels in 12), the bit corresponding to the column with a small weight in the irregular LDPC matrix H is transmitted using a communication channel having a good communication state among the plurality of communication channels. In addition, the bit corresponding to the column having a large weight in the irregular LDPC matrix H may be transmitted using another communication path.

  In the aspect shown in FIG. 12, before transmitting transmission target information (transmission word), transmission apparatus 10 first transmits known signals (pilot signals) via M communication paths. The known signals respectively transmitted from the transmitting device 10 via the M communication paths are received by the receiving device 20, and the receiving device 20 receives the reception quality of the known signals received via the respective communication paths (for example, the received signal). The communication state of each channel is estimated based on the strength, signal power to noise power ratio (SNR), signal power to interference + noise power ratio (SINR), spatial correlation, etc.) The state estimation result is notified to the transmission device 10 as channel state information. Thereby, the transmission apparatus 10 can detect the communication state of M communication paths. When the M communication channels are communication channels whose communication state fluctuates with time, the transmission device 10 periodically transmits a known signal, and the reception device 20 communicates every time a known signal is received. It is desirable that the communication state of the M communication paths is periodically confirmed by notifying the estimation result of the communication state of the communication path by the path state information.

  When the state of the communication path is detected, the modulator 62 of the transmission apparatus 10 corresponds to the communication path used for transmitting each bit of the encoded information input from the LDPC encoder 12 to a column having a small weight in the irregular LDPC matrix. A bit having a relatively good communication state is assigned to the bit to be transmitted, and a bit corresponding to a column having a large weight in the irregular LDPC matrix is a bit corresponding to a column having a small weight in the irregular LDPC matrix. It is determined that a communication channel having a communication state worse than that of the communication channel allocated to (1) is allocated. The modulator 62 rearranges and modulates the input encoded information according to the determined communication channel assignment, and the serial-parallel converter 64 receives the input according to the communication channel assignment to each bit of the encoded information. By appropriately selecting a communication channel for transmitting the transmission signal, the transmission signal is distributed and transmitted to a plurality of different communication channels.

  When a signal distributed and transmitted to a plurality of communication paths is received from the transmission device 10, the demodulator 68 of the reception device 20 demodulates each received signal input via the parallel-serial conversion unit 66, and then rearranges the received signals. To generate encoded information (received word) and input it to the LDPC decoder 26. As a result, the transmitted word (estimated result) is output from the LDPC decoder 26.

  In the above aspect, when transmitting encoded information, bits corresponding to a column having a small weight on the irregular LDPC matrix H among the bits of the encoded information are transmitted using a communication channel having a relatively good communication state. Therefore, the encoded information (received word) output from the demodulator 68 of the receiving device 20 is similar to the mode in which a large transmission power is assigned to the bit corresponding to the column having a small weight on the irregular LDPC matrix H. , The probability that the bit value corresponding to the column with a small weight on the irregular LDPC matrix H is changed can be reduced, and the BER characteristic at the time of decoding can be improved.

  Next, a simulation performed by the present inventor in order to confirm the effect when the power allocation method according to the present invention is applied will be described.

  In this simulation, as a power allocation method according to the present invention, a method is applied in which the power allocated to each bit of encoded information is inversely proportional to the square of the weight of the corresponding column of the irregular LDPC matrix. An example in which no allocation is performed (referred to as “no power allocation”), an “inverse power allocation method” in which small power is allocated to a bit corresponding to a column having a small weight in an irregular LDPC matrix, and allocation to each bit of encoded information A “random power allocation method” that randomly changes the power was used. Table 1 below shows an example of the column weight of the irregular LDPC matrix and the ratio of the power allocated to the corresponding bits in each of the above methods (excluding the random power allocation method).

  Then, an irregular LDPC code having a code length N = 512 and an information bit number K = 195 is used as an LDPC code used for encoding transmission target information, a modulation method is set to BPSK modulation, and a decoding algorithm is set to a BP algorithm. (However, the maximum number of repetitions = 100), an additive white Gaussian noise (AWGN) channel is applied to the channel model, and the calculation is performed for each of the above methods, whereby the average information bit energy to noise power ratio Eb / The relationship between N0 and the bit error rate BER was obtained (results are shown in FIG. 13). In this simulation, the transmission power in each method is normalized to 1, and the calculation is performed so that the average value of the transmission power in each method including the random power allocation method is 1.

  As is apparent from FIG. 13, when the power allocation method according to the present invention is applied, the bit error rate BER is improved over the other methods when the average information bit energy-to-noise power ratio Eb / N0 is in the range of 4-6. Therefore, it can be understood that the power allocation method according to the present invention is effective in improving the BER characteristics.

It is a block diagram which shows schematic structure of the transmitter which concerns on this embodiment. It is a block diagram which shows an example of schematic structure of the receiver which concerns on this embodiment. (A) is a normal BPSK modulation, and (B) is a diagram showing an example of BPSK modulation to which the present invention is applied. It is explanatory drawing which shows the notation in description of a decoding process. It is a conceptual diagram which shows an example of propagation of the probability information from a bit node to a check node. It is a conceptual diagram which shows an example of propagation of the probability information from a check node to a bit node. It is a flowchart which shows the content of the decoding process which concerns on this embodiment. It is a conceptual diagram which shows an example of propagation of the probability information from the check node m1 to the bit node n2. It is a conceptual diagram which shows an example of propagation of the probability information from the bit node n1 to the check node m2. When an OFDM modulation system is applied as a modulation system, (A) is a block diagram showing a schematic configuration of a transmission apparatus and (B) is a reception apparatus. It is a block diagram which shows the other example of schematic structure of a receiver. It is a block diagram which shows schematic structure of the transmitter which concerns on other embodiment, and a receiver. It is a diagram which shows the result of experiment. It is a conceptual diagram which shows an example of a Factor graph.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Transmitter 12 Encoder 14 Modulator 16 Antenna 20 Receiver 22 Antenna 24 Demodulator 26 Decoder 28 Correlator 30 Mapping / power allocation unit

Claims (9)

In encoding and transmitting information to be transmitted using a low density parity check code defined by a low density parity check matrix in which the weight of each row and each column is not constant,
Of each bit of the encoded information obtained by encoding the information to be transmitted, a bit corresponding to a column with a small weight in the low density parity check matrix is a column with a large weight in the low density parity check matrix. A method of transmitting information, wherein the encoded information is transmitted so that the probability of receiving without error on the receiving side is higher than bits corresponding to.   Of each bit of the encoded information, a bit corresponding to a column having a small weight in the low-density parity check matrix is more error-free on a receiving side than a bit corresponding to a column having a large weight in the low-density parity check matrix. The information according to claim 1, wherein the encoded information is transmitted by increasing a transmission power allocated to a bit corresponding to the column having a small weight so that the probability of reception is high. Transmission method.   3. The information transmission method according to claim 2, wherein transmission power allocated to a bit corresponding to a column having a large weight in the low-density parity check matrix among the bits of the encoded information is reduced.   4. The information transmission method according to claim 2, wherein transmission power allocated to each bit of the encoded information is inversely proportional to a square of a weight of a corresponding column in the low density parity check matrix.   The transmission power allocated to each bit of the encoded information is changed by partially changing the amplitude of a carrier wave modulated in accordance with the encoded information. Item 4. The information transmission method according to Item 3.   Changing the transmission power allocated to each bit of the encoded information is modulated according to the bit corresponding to the column having the smaller weight in the transmission signal obtained by modulating the carrier wave according to the encoded information. 4. The information transmission method according to claim 2, wherein the information transmission method is performed by transmitting the determined portion a plurality of times. Of each bit of the encoded information, a bit corresponding to a column having a small weight in the low-density parity check matrix is more error-free on a receiving side than a bit corresponding to a column having a large weight in the low-density parity check matrix. Transmitting the encoded information so that the probability of being received is high,
A bit corresponding to a column having a small weight in the low density parity check matrix is a channel having a good communication state among a plurality of channels separated by at least one of a frequency domain, a time domain, a spatial domain, and a code domain. And transmitting a bit corresponding to a column having a large weight in the low density parity check matrix to a communication channel different from a transmission path used for transmitting a bit corresponding to a column having a small weight in the low density parity check matrix The information transmission method according to claim 1, wherein the information transmission method is performed by transmitting the information.   Each of the plurality of communication paths is transmitted based on a reception state on the receiving side of the known signal transmitted through each of the plurality of communication paths, each of the known signals being transmitted using the plurality of communication paths. The information transmission method according to claim 7, wherein a state is detected. An information transmitting apparatus that encodes and transmits information to be transmitted using a low density parity check code defined by a low density parity check matrix in which the weight of each row and each column is not constant,
Of each bit of the encoded information obtained by encoding the information to be transmitted, a bit corresponding to a column with a small weight in the low density parity check matrix is a column with a large weight in the low density parity check matrix. An information transmitting apparatus comprising: transmission control means for transmitting the encoded information so that the probability of receiving without error on the receiving side is higher than bits corresponding to.

Images