CN115426014B - Underwater sound MIMO communication method based on unitary space-time code modulation - Google Patents

Abstract

The invention provides an underwater acoustic MIMO communication method based on unitary space-time coding modulation, which belongs to the technical field of underwater communication. The present invention uses unitary space-time modulation that does not require CSI to achieve high reliability of communication; it uses MIMO technology to obtain the power gain and multipath diversity gain provided by large-scale arrays, significantly improve the power efficiency and spectrum efficiency of the system, and improve the system transmission rate and transmission distance. Using LDPC coding can reduce the hardware implementation complexity without reducing coding performance. The proposed joint iterative algorithm between the unitary space-time demodulator and the LDPC decoder can effectively improve system performance. The invention has the characteristics of ultra-low speed, anti-interference and high reliability.

Description

An underwater acoustic MIMO communication method based on unitary space-time coding modulation

Technical field

The present invention relates to the technical field of underwater acoustic wireless communication, and in particular to an underwater acoustic MIMO communication method based on unitary space-time coding modulation.

Background technique

Multipath propagation, Doppler frequency shift and other phenomena widely exist in underwater acoustic communication channels, which can cause serious interference to communication quality. Using multiple input multiple output (MIMO) technology, under the condition of limited underwater spectrum resources, the spectrum efficiency and power efficiency of the underwater acoustic communication system can be effectively improved, thereby effectively improving the data rate and communication link reliability. In addition, in order to effectively overcome interference, Space Time Coding (STC), which can fully obtain space diversity gain and coding gain in MIMO technology, has been a research hotspot.

One of the main challenges of MIMO is the availability of accurate channel state information (CSI) for coherent detection and/or transmitter precoding. On the one hand, as the number of antennas or transducers at the transceiver increases, the channel estimation becomes increasingly difficult. It becomes increasingly difficult, especially in some fast fading environments where the environment has changed before channel estimation is completed. In addition, sending training sequences for channel estimation results in a waste of channel bandwidth, greatly reduces frequency band utilization, and makes the already limited underwater acoustic channel bandwidth even tighter. If the traditional space-time coding scheme cannot accurately obtain CSI, it will lead to communication failure and bring unpredictable consequences in some fields.

Therefore, there is an urgent need to study space-time codes that do not require channel information under fast fading conditions. The use of unitary space-time coded modulation (USTM) without CSI can meet various requirements such as high reliability of underwater wireless communications and has extremely high reliability. great research value.

Contents of the invention

The purpose of the present invention is to provide an underwater acoustic MIMO communication method based on unitary space-time coding modulation to make up for the shortcomings of the existing technology.

In order to achieve the above objects, the specific technical solutions adopted by the present invention are:

An underwater acoustic MIMO communication method based on unitary space-time coding modulation, including the following steps:

S1: Transmit a signal, convert the signal into binary information bits, and then perform LDPC encoding to map the encoded data to a unitary matrix constellation set;

S2: Send the mapped unitary matrix signal to the transmitting hydroacoustic transducer and reach the receiving end through the hydroacoustic channel. The receiving end uses the USTM demodulator to calculate the a priori information transmitted in the received signal sequence and LDPC decoder. Log-likelihood ratio soft information under μ conditions; the receiving end is equipped with N receiving underwater acoustic transducer arrays, the USTM demodulator calculates the received signal sequence and the a priori information μ transmitted by the LDPC decoder (first external Iteratively set it to the posterior log-likelihood ratio under the condition of a sparse matrix with all elements equal to 0), and extract the prior information μ from the posterior log-likelihood information to generate information μ';

S3: The log-likelihood ratio soft information is calculated to obtain external information and provided to the LDPC decoder; the output result of the LDPC decoder is fed back to the USTM demodulator, and an external iteration relationship is introduced to construct an Joint iterative decoding. Joint iterative decoding includes the USTM demodulator and the LDPC decoder. After continuous iterative decoding, the decoding result is finally output.

The μ' obtained by S2 is provided to the LDPC decoder as external information. The LDPC decoder is composed of a variable node decoder and a check node decoder. The decoding algorithm adopts the logarithmic domain belief propagation (LLR BP) algorithm. ; In order to make full use of the high reliability output results of the LDPC code, the output results of the LDPC code are fed back to the USTM demodulator, and a joint iterative decoding algorithm and joint decoding module are constructed by introducing this external iterative relationship. It includes USTM demodulator and LDPC decoder.

Among them, on the one hand, the LDPC decoder receives external information as its prior information to perform internal iteration decoding between variable nodes and check nodes; on the other hand, it feeds back the result information of the internal iteration to the USTM demodulator for the next time. outer iteration.

Further, the S1 includes:

S1-1: k binary information _{bits expressed} as ={c ₁ ,…,c _n }; where k represents the length of information bits, and the length of information bits after LDPC encoding is expressed as n;

S1-2: Take out q bits from the n-long codeword sequence in sequence (K=n/q, that is, K groups), and map these q bits according to a certain mapping rule in each time block T. into a T×M Φ _l matrix signal, Φ _l is any one of L=2 ^q unitary matrix constellation sets Ω={Φ ₁ ,...,Φ _L }, l=1,...L, M is The number of transmitting transducers, the mapping from q bits to a Φ _l matrix signal point is Gray mapping.

Furthermore, in S1, the unitary matrix constellation set adopts a systematic and simple construction method: Φ _l =Θ ^l-1 Φ ₁ , Φ ₁ is a T×M dimensional unitary matrix, and its composition can be derived from a T×T dimensional DFT Select any M columns in the matrix, Θ is a T×T dimensional diagonal matrix, and its Lth power is the T×T dimensional unit matrix I _T .

The S2 is specifically:

S2-1: The received signal Y passing through the underwater acoustic channel corresponding to the transmitted signal Φ _l , that is, the unitary matrix constellation set, expressed as the following formula:

Here H represents the channel fading coefficient, and W represents additive Gaussian white noise. In order to effectively model the real underwater acoustic environment, based on the characteristics of the underwater acoustic channel, combined with the characteristics of USTM, considering the Doppler effect and the influence of variability in the underwater acoustic channel, based on the BELLHOP model, an underwater acoustic channel suitable for USTM was constructed model, H of the underwater acoustic channel environment is obtained through modeling and simulation; that is, the amplitude A, phase φ and other parameters are obtained through simulation of the underwater acoustic channel model. The channel fading coefficient of the time block T on each channel can be modeled as a complex random variable:

S2-2: The length of codeword c is n, divide n into K groups, each group has q bits, that is, n=Kq; K is the Kth group, m is the mth bit of the Kth group, where K=1,…, K and m=1,…,q,f(K) represent a certain mapping rule, mapping each group c _K to the corresponding unitary matrix Φ _(f(K)) , Φ _(f(K)) ∈Ω, Ω represents the unitary constellation set. For a codeword c, all received signals are represented as Y. Then the posterior probability log-likelihood ratio Λ(c(d)) for the d-th bit c(d) (d=1,2,...,n) of codeword c can be expressed as:

S2-3: Assume that the d-th bit of codeword c is the m-th bit of the K-th group, expressed as c _K ^m (d), and c _K represents all the q-bits of the K-th group. It is assumed that the emission probabilities of each signal matrix are equal. , then Λ(c(d)) can be further written as:

S2-3: Assume that the d-th bit of codeword c is the m-th bit of the K-th group, expressed as c _K ^m (d), and c _K represents all the q-bits of the K-th group. It is assumed that the emission probabilities of each signal matrix are equal. , then Λ(c(d)) can be further written as:

S2-4: In unitary space-time modulation, under the premise that the transmit signal matrix is known to be X, the conditional probability density function of the received signal matrix Y is:

ξ represents the conjugate transpose of the matrix. Simplifying the conditional probability density function of the received signal matrix Y in the above formula is:

Substituting the conditional probability density function of Y into Λ(c(d)), Λ(c(d)) can be further expressed as:

The input prior information is subtracted from the obtained posterior LLR, and the external information is obtained and output to the LDPC decoder. That is, the external information value can be given by the following formula:

.

Further, the S3 is specifically:

First, the symbols of the LDPC decoding algorithm are explained: i is the check node, j is the variable node, z is the current iteration number; P _j (1) is the posterior probability of receiving the bit c _j =1 sent by the sender of y _j , P _j (0) is the posterior probability of receiving bit c _j =0 sent by the sender of y _j , q _ji ^z(b) is the external information transmitted by node j to node i at the z-th iteration, b= 0,1. r _ij ^z(b) is the external information transmitted by node i to node j in the z-th iteration, Q _j ^z (b) is the hard decision message in the z-th iteration; C(j) indicates that it is connected to j The set of check nodes, C(j)={i:h _ij =1}, V(i) represents the set of variable nodes connected to i, V(i)={j:h _ij =1}; C(j )\i represents the set of check nodes connected to j except i, C(j)\i={ u:h _ij =1,u≠i}; V(i)\j represents the set of check nodes connected to i except j The set of variable nodes, V(i)\j={ u:h _ij =1,u≠j}.

S3-1: Initialization: Calculate the initial prior probability information of each variable node and the connected check node, expressed as:

Update of information from check nodes to variable nodes: Calculate and update the information passed by all check nodes to connected variable nodes, expressed as:

Update of information from variable nodes to check nodes: Calculate and update the information from all variable nodes to the connected check nodes, expressed as:

LLR sum:

Make a decision on Γ ^(z) (Q _j ), if Γ ^(z) (Q _j )<0, then x'=1; if Γ ^(z) (Q _j )≥0, then x'=0, and finally form Codeword x=[x ₁ ,x ₂ ,…,x _n ];

S3-2: Judgment: If the check matrix B and the final codeword x satisfy Bx ^T =0, the decoding ends and the entire decoding process ends. The codeword x=[x ₁ ,x ₂ ,…,x _n ] is The final decoding result;

S3-3: Output the ^decoded message: Select the first 1 to k items in the codeword After the inner iteration process ends, the soft information value output by the inner iteration is fed back to the USTM demodulator, and the next outer iteration is performed. The entire decoding process is terminated until the maximum number of outer iterations is reached.

The advantages and beneficial effects of the present invention are:

The present invention uses unitary space-time modulation that does not require CSI to achieve high reliability of communication; it uses MIMO technology to obtain the power gain and multipath diversity gain provided by large-scale arrays, significantly improve the power efficiency and spectrum efficiency of the system, and improve the system transmission rate and transmission distance. Using LDPC coding can reduce the hardware implementation complexity without reducing coding performance. The proposed joint iterative algorithm between the unitary space-time demodulator and the LDPC decoder can effectively improve system performance. The invention has the characteristics of ultra-low speed, anti-interference and high reliability.

The joint iterative decoding scheme proposed by the present invention makes full use of the high-reliability soft information fed back by the LDPC decoder, and significantly improves the bit error performance of the LDPC-USTM acoustic cascade space-time coding system. The present invention provides a system based on unitary space-time coding modulation that can not only solve the problem of instability and unreliability of traditional space-time coding in MIMO underwater acoustic communication systems, but is also feasible in hardware and can further improve the reliability of underwater acoustic communication. Hydroacoustic communication methods.

Description of drawings

Figure 1 is a complete signal sending and receiving flow chart;

Figure 2 is the LDPC-USTM simulation flow chart;

Figure 3 is the LDPC decoding flow chart;

Figure 4 is a comparison chart of the USTM underwater acoustic communication bit error rate curves using simple cascaded LDPC coding, the proposed iterative LDPC coding and uncoding.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings:

Example 1:

An underwater acoustic MIMO communication method based on unitary space-time coding modulation. The basic flow of the method is shown in Figure 1, which specifically includes the following steps:

S1: Transmit a signal, convert the signal into binary information bits, and then perform LDPC encoding;

_The LDPC encoding process is as follows: k binary information bits expressed _as c={c ₁ ,…,c _n }. Here k represents the length of information bits, and n represents the length of information bits after LDPC encoding. Sequentially take out q bits from the n-long codeword sequence (K=n/q, that is, K groups), and in each time block T, map these q bits into a T× Φ _l matrix signal of M, where Φ _l is any one of L=2 ^q unitary matrix constellation sets Ω={Φ ₁ ,...,Φ _L }, l=1,...L, M is the number of transmitting terminals , the mapping from q bits to a Φ _l matrix signal point can be Gray mapping.

S2: The transmitted signal Φ _l reaches the receiving end after passing through the underwater acoustic channel. The receiving end is equipped with N receivers; the USTM demodulator calculates the received signal sequence and the a priori information μ transmitted by the LDPC decoder (the first outer iteration sets it is the posterior log-likelihood ratio under the condition of a sparse matrix with all elements equal to 0), and the prior information μ is extracted from the posterior log-likelihood information to generate information μ'.

S3: μ' is provided to the LDPC decoder as external information. The LDPC decoder consists of a variable node decoder and a check node decoder. The decoding algorithm uses the logarithmic domain belief propagation algorithm (LLR BP). The decoded information is decoded externally between the USTM demodulator and the LDPC decoder, and internally iterated inside the LDPC decoder. After a certain number of iterations, the final information is decoded and judged, and the judgment result is finally output.

As shown in Figure 2, based on the underwater acoustic MIMO communication method, the underwater acoustic channel is modeled and specific experiments are conducted, including:

1. First, model the underwater acoustic channel environment suitable for USTM and restore the real underwater acoustic environment.

2. The specific process for the receiving end to obtain the log-likelihood ratio soft information is:

The length of the codeword c is n, and n is divided into K groups, each group has q bits, that is, n=Kq; K is the Kth group, m is the mth bit of the Kth group, where K=1,...,K and m= 1,…,q, f(K) represents a certain mapping rule, mapping each group c _K to the corresponding unitary matrix Φ _(f(K)) , Φ _(f(K)) ∈Ω, Ω represents the unitary constellation set , for a codeword c, all received signals are represented as Y. Then the posterior probability log-likelihood ratio Λ(c(d)) for the d-th bit c(d) (d=1,2,...,n) of codeword c can be expressed as:

Assume that the d-th bit of codeword c is the m-th bit of the K-th group, expressed as c _K ^m (d), c _K represents all q-bits of the K-th group, assuming that the emission probabilities of each signal matrix are equal, then Λ( c(d)) can be further written as:

In unitary space-time modulation, under the premise that the transmitted signal matrix is known as X, the conditional probability density function of the received signal matrix Y is:

Represents the conjugate transpose of the matrix, and simplifies the conditional probability density function of the received signal matrix Y to obtain:

Substituting the conditional probability density function of Y into Λ(c(d)), Λ(c(d)) can be further expressed as:

The input prior information is subtracted from Λ(c(d)) to obtain the external information and output to the LDPC decoder. That is, the external information value can be given by the following formula:

;

3. The specific process of LDPC decoding is (shown in Figure 3):

Initialization: Calculate the initial information of each variable node and the connected check node, expressed as:

Update of information from check nodes to variable nodes: Calculate and update the information passed by all check nodes to connected variable nodes, expressed as:

Update of information from variable nodes to check nodes: Calculate and update the information from all variable nodes to the connected check nodes, expressed as:

LLR sum:

Make a decision on Γ ^(z) (Q _j ), if Γ ^(z) (Q _j )<0, then x'=1; if Γ ^(z) (Q _j )≥0, then x'=0, and finally form Codeword x.

Judgment: If the check matrix B and the final codeword x satisfy Bx ^T =0, the decoding ends and the entire decoding process ends. The codeword x=[x ₁ ,x ₂ ,…,x _n ] is the final decoding result. Output the decoded message: Select the first 1 to k items in the codeword x to be the decoded message. Otherwise, repeat the above inner iteration steps until Bx ^T =0 is satisfied or the maximum number of inner iterations is exceeded. At this time, the inner iteration process ends, and the soft information value output by the inner iteration is fed back to the USTM demodulator, and the next outer iteration is performed until When the maximum number of outer iterations is reached, the entire decoding process is terminated. The proposed joint iterative decoding scheme makes full use of the high-reliability soft information fed back by the LDPC decoder, and significantly improves the bit error performance of the LDPC-USTM acoustic cascade space-time coding system.

Example 2:

Simulation conditions: number of transmitting transducers M=2, number of receiving transducers N=1, channel coherence time T=4ms, unitary space-time symbol rate R=1, LDPC code rate 1/5, code length 1280bit, external iteration 1 time, 10 iterations. Underwater acoustic channel parameters: water depth 100m, transmitter depth 50m, receiver depth 50m, transmission distance 2km, carrier frequency 10KHz.

When the underwater acoustic communication method adopts the unitary space-time coding modulation of the present invention, a comparison of the bit error rate curves is obtained, as shown in Figure 4. It can be seen from Figure 4 that as the signal-to-noise ratio increases, the number of error bits gradually decreases and the bit error rate curve gradually converges; compared with the uncoded unitary space-time modulation system, when the bit error rate is 10 ^-4 , the unitary space-time system using LDPC coding can provide a signal-to-noise ratio gain of about 18dB. The proposed joint iterative LDPC-USTM scheme has a gain of about 0.5dB compared with the simple cascade LDPC-USTM scheme.

It can be seen from the comparison results that under the underwater acoustic channel, the unitary space-time modulation method using channel coding can bring more signal-to-noise ratio gain than the uncoded unitary space-time method. The joint iterative LDPC- The USTM decoding scheme is more reliable and has better decoding effect than the traditional simple cascaded LDPC-USTM scheme.

In summary, the present invention considers the need for high reliability of underwater communications in some cases. If a technology that requires CSI is used, once the channel estimation cannot complete the task, accurate CSI cannot be obtained, and the entire frame fails, which will cause Serious communication problems arise. Applying unitary space-time coded modulation in underwater acoustic communications has certain research and application value.

Claims (2)

1. An underwater acoustic MIMO communication method based on unitary space-time coding modulation, which is characterized by including the following steps: S1: Transmit a signal, convert the signal into binary information bits, and then perform LDPC encoding to map the encoded data to a unitary matrix constellation set; the S1 includes: S1-1: k binary information _{bits expressed} as ={c ₁ ,…,c _n }; where k represents the length of information bits, and the length of information bits after LDPC encoding is expressed as n; S1-2: Take out q bits from the n-long codeword sequence in turn, and map these q bits into a T×M Φ _l matrix signal according to a certain mapping rule in each time block T, Φ _l is any one of L=2 ^q unitary matrix constellation sets Ω={Φ ₁ ,...,Φ _L }, l=1,...L, M is the number of transmitting transducers, from q bits to a Φ _lThe mapping of matrix signal points is Gray mapping; S2: Send the mapped unitary matrix signal to the transmitting hydroacoustic transducer and reach the receiving end through the hydroacoustic channel. The receiving end uses the USTM demodulator to calculate the a priori information transmitted in the received signal sequence and LDPC decoder. For log-likelihood ratio soft information under μ conditions, the prior information μ is extracted from the posterior log-likelihood information to generate information μ', and μ' is provided to the LDPC decoder as external information; the S2 is specifically: S2-1: The received signal Y passing through the underwater acoustic channel corresponding to the transmitted signal Φ _l forms a unitary matrix constellation set, expressed as the following formula: H represents the channel fading coefficient, W represents the additive Gaussian white noise. According to the characteristics of the underwater acoustic channel and combined with the characteristics of USTM, considering the Doppler effect and the influence of variability in the underwater acoustic channel, based on the BELLHOP model, a suitable USTM's underwater acoustic channel model, through modeling and simulation, shows that the underwater acoustic channel environment is modeled as a complex random variable: S2-2: The length of codeword c is n, divide n into Q groups, each group has q bits, that is, n=Qq; K is the Kth group, m is the mth bit of the Kth group, where K=1,..., Q and m=1,...,q,f(K) represent a certain mapping rule, mapping each group c _K to the corresponding unitary matrix Φ _(f(K)) ,Φ _(f(K)) ∈Ω, for For a codeword c, all received signals are represented as Y; then the posterior probability log-likelihood ratio Λ(c(d)) of the d-th bit c(d) of the codeword c is expressed as: S2-3: Assume that the d-th bit of codeword c is the m-th bit of the K-th group, expressed as c _K ^m (d), and c _K represents all the q-bits of the K-th group. It is assumed that the emission probabilities of each signal matrix are equal. , then Λ(c(d)) is further written as: S2-4: In unitary space-time modulation, under the premise that the transmitted signal matrix is known as X, the conditional probability density function of the received signal Y is: ξ represents the conjugate transpose of the matrix. Simplifying the conditional probability density function of the received signal matrix Y in the above formula is: Substituting the conditional probability density function of Y into Λ(c(d)), Λ(c(d)) can be further expressed as: Subtracting the input prior information from Λ(c(d)), the external information is output to the LDPC decoder; the external information value is given by the following formula: μ′=Λ(c(d))-μ; S3: The log-likelihood ratio soft information is calculated to obtain external information and provided to the LDPC decoder; the output result of the LDPC decoder is fed back to the USTM demodulator, and an external iteration relationship is introduced to construct an Joint iterative decoding. Joint iterative decoding includes the USTM demodulator and the LDPC decoder. After continuous iterative decoding, the decoding result is finally output; the USTM demodulator calculates the signal sequence received from the channel and Receive the information fed back by the LDPC decoder, use the received information to calculate the external information of each transmitted signal, and then send it to the LDPC decoder for decoding; the LDPC decoder receives the external information as its prior information Perform inner iteration decoding between variable nodes and check nodes, and feed back the result information of the inner iteration to the USTM demodulator for the next outer iteration; the S3 is specifically: S3-1: Initialization: Calculate the initial prior probability information of each variable node and the connected check node, expressed as: Γ ⁽⁰⁾ (q _ji )=μ Update of information from check nodes to variable nodes: Calculate and update the information passed by all check nodes to connected variable nodes, expressed as: Update of information from variable nodes to check nodes: Calculate and update the information from all variable nodes to the connected check nodes, expressed as: L ^(z) (q _ji )=μ′+∑ _u∈C(j)\i L ^(z) (r _uj ) LLR sum: L ^(z) (Q _j )=μ′+∑ _i∈V(j) L ^(z) (r _ij ) Make a decision on Γ ^(z) (Q _j ). If Γ ^(z) (Q _j )<0, then x'=1; if Γ ^(z) (Q _j )≥0, then x'=0, and finally form Codeword x=[x ₁ ,x ₂ ,…,x _n ]; S3-2: If the check matrix B and the final codeword x satisfy Bx ^T = 0, the decoding ends and the entire decoding process ends. The codeword x = [x ₁ , x ₂ ,..., x _n ] is the final Decoding results; S3-3: Output the decoded message: Select the first 1 to k items in the codeword x as the decoded message; otherwise, repeat the above inner iteration steps until Bx ^T = 0 or exceed the maximum number of inner iterations. End the inner iteration process, feed back the soft information value output by the inner iteration to the USTM demodulator, and perform the next outer iteration until the maximum number of outer iterations is reached, then the entire decoding process is terminated; Among them, the explanation in the above formula is: i is the check node, j is the variable node, z is the current iteration number; P _j (1) is the posterior probability of receiving the bit c _j =1 sent by the sender of y _j , P _j (0) is the posterior probability of receiving the bit c _j =0 sent by the sender of y _j , q _ji ^z(b) is the external information transmitted by node j to node i at the z-th iteration, b=0 ,1; r _ij ^z(b) is the external information transmitted by node i to node j in the z-th iteration, Q _j ^z(b) is the hard decision message in the z-th iteration; C(j) represents The set of check nodes connected to j, C(j)={i:h _ij =1}, V(i) represents the set of variable nodes connected to i, V(i)={j:h _ij =1}; C(j)\i represents the set of check nodes connected to j except i, C(j)\i＝{u:h _ij ＝1,u≠i}; V(i)\j represents except j The set of variable nodes connected to i, V(i)\j={u:h _ij =1,u≠j}. 2. The underwater acoustic MIMO communication method according to claim 1, characterized in that the construction method of the unitary matrix constellation set is: Φ _l = Θ ^l-1 Φ ₁ , Φ ₁ is a T×M dimensional unitary matrix, from Select M columns in the T×T dimensional DFT matrix, Θ is a T×T dimensional diagonal matrix, and its Lth power is the T×T dimensional unit matrix I _T .