A distributed MIMO single-carrier underwater acoustic communication joint channel estimation and symbol detection method

Abstract

The application belongs to the field of distributed MIMO underwater acoustic communication, and discloses a kind of distributed MIMO single carrier underwater acoustic communication joint channel estimation and symbol detection method, comprising: constructing system model;Channel model is constructed;Guiding cost function;Initialization parameter;Parameter is updated;Estimate symbol;Estimate channel tap;Estimate noise variance;Estimate first-order Markov coefficient;Estimate channel noise variance.The application uses the method based on sub-block, can overcome the problem of non-cooperative arbitrary sending under distributed MIMO scene, inter-channel interference and near-far effect, and will establish first-order Markov model of channel, realizes reliable distributed MIMO underwater acoustic communication under slow time-varying underwater acoustic channel.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of distributed MIMO underwater acoustic communication, and particularly to a joint channel estimation and symbol detection method for distributed MIMO single-carrier underwater acoustic communication. BACKGROUND

[0002] With the further research and development of the ocean, the demand for underwater high-speed wireless networking communication technology is becoming more and more urgent. Underwater acoustic communication has natural advantages in ocean environment propagation and can propagate over a longer distance. However, underwater acoustic communication technology also needs to achieve higher communication rates and adapt to more underwater application scenarios, such as multi-node networking. Distributed MIMO single-carrier underwater acoustic communication can break the limitation that each node needs to communicate in turn in networking communication, allowing multiple transmitting nodes to communicate to the receiving node at the same time, thereby improving the throughput of networking communication between nodes and improving the timeliness of node communication.

[0003] In distributed MIMO underwater acoustic communication, multiple different transmitting transducers are distributed on different nodes and are independently controlled by different transmitters. The transmitters cannot know the transmission data of other transmitters. The receiving end receives signals from multiple platforms through multiple hydrophones and demodulates the signals. The signals from the transmitting transducers to the receiving end hydrophones experience different large-scale fading and small-scale fading, thus producing near-far effects and experiencing different slow time-varying channels. In addition, the signals transmitted by different transmitting transducers will collide at the receiving end, thus existing different staggered times and inter-channel interference.

[0004] Because of the different staggered times, the traditional RLS-DFE and MIMO-DFE demodulation methods for traditional MIMO are no longer applicable to the distributed MIMO scenario. Therefore, a block processing-based method can be used to divide the received signals into smaller sub-blocks and jointly estimate the channel and symbol in the sub-blocks. The present application proposes a joint channel estimation and symbol detection method for distributed MIMO single-carrier underwater acoustic communication from the perspective of maximum a posteriori probability estimation, and uses the block coordinate descent method to realize the optimization estimation of the channel and symbol, so that better channel estimation and symbol detection performance can be obtained in the distributed MIMO slow time-varying environment. SUMMARY

[0005] The present application aims to provide a joint channel estimation and symbol detection method for distributed MIMO single-carrier underwater acoustic communication to solve the problems in the background art.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions.

[0007] A joint channel estimation and symbol detection method for distributed MIMO single-carrier underwater acoustic communication, the method comprising the following steps:

[0008] S1: constructing a distributed MIMO single-carrier underwater acoustic communication system, containing M transducers and N hydrophones;

[0009] S2: constructing a channel model, the channel between sub-blocks is represented by a first-order Markov model;

[0010] S3: according to the distributed MIMO single-carrier underwater acoustic communication system and the channel model, a cost function of the distributed MIMO single-carrier underwater acoustic communication system is derived from the perspective of maximum a posteriori probability;

[0011] S4: initializing the estimated value of the channel tap of the 0th sub-block in the channel model, the noise variance, the first-order Markov coefficient and the noise variance of the channel;

[0012] S5: in the kth sub-block, firstly, the estimated value of the channel tap, the estimated value of the noise variance, the estimated value of the first-order Markov coefficient and the estimated value of the channel noise variance are kept consistent with the k-1th sub-block;

[0013] Then the symbol of the kth sub-block is estimated to obtain its estimated value, and then the estimated value of the channel tap, the estimated value of the noise variance, the estimated value of the first-order Markov coefficient and the estimated value of the channel noise variance are updated in turn;

[0014] S6: using the block coordinate descent method to iterate and repeat S5 until the number of iterations reaches the maximum number of iterations to finally realize joint channel estimation and symbol detection.

[0015] Further, the S1 comprises: the M transducers are distributed on M different transmitting nodes, each of which is independently controlled, and each transmitter cannot know the transmission data of other transmitters; the N hydrophones are distributed on the same receiving end to receive signals from multiple transmitting nodes and demodulate; the M different transmitting nodes have a speed close to the receiving end, and the distance between the M transducers and the receiving end is arbitrary, and the transmission time of the M transducers is arbitrary.

[0016] Further, the initialization of the estimated value of the channel tap of the 0th sub-block in the S4 comprises:

[0017] S4.1: constructing a Toeplitz matrix of the training symbol transmitted by the mth transducer;

[0018] S4.2: intercepting the part of the training symbol vector of the mth transducer corresponding to the data received by the nth hydrophone;

[0019] S4.3: The channel tap estimate value of the 0th sub-block between the mth transducer and the nth hydrophone is expressed by least square method, that is, the pseudo-inverse of the Toeplitz matrix of the training symbols of the data sent by the mth transducer is multiplied by the part of the corresponding mth transducer training symbol vector in the data received by the nth hydrophone.

[0020] Further, the S5 specifically comprises:

[0021] S5.1: In the kth sub-block, the estimate value of the channel tap, the estimate value of the noise variance, and the estimate value of the first-order Markov coefficient are first kept consistent with the k-1th sub-block;

[0022] S5.2: The symbol of the kth sub-block is estimated to obtain its estimate value, comprising: in the kth sub-block, the symbol of the kth sub-block is estimated to obtain its estimate value according to the received signal, the estimate value of the channel tap, the estimate value of the noise variance, and the variance of the symbol;

[0023] S5.3: The estimate value of the channel tap is updated, comprising: in the kth sub-block, the estimate value of the channel is updated according to the received signal, the estimate value of the symbol, the estimate value of the noise variance, the estimate value of the first-order Markov coefficient, and the estimate value of the channel noise variance;

[0024] S5.4: The estimate value of the noise variance is updated, comprising: in the kth sub-block, the estimate value of the noise variance is updated according to the received signal, the estimate value of the symbol, the estimate value of the channel tap, and the length of the sub-block;

[0025] S5.5: The estimate value of the first-order Markov coefficient is updated, comprising: in the kth sub-block, the estimate value of the first-order Markov coefficient is updated according to the estimate value of the channel tap of the current kth sub-block and the estimate value of the channel tap of the k-1th sub-block;

[0026] S5.6: The estimate value of the channel noise variance is updated, comprising: in the kth sub-block, the estimate value of the channel noise variance is updated according to the estimate value of the channel tap of the current kth sub-block, the estimate value of the channel tap of the k-1th sub-block, the estimate value of the first-order Markov coefficient of the k-1th sub-block, and the length of the channel.

[0027] Further, the S5.2 specifically comprises: according to the cost function of the kth sub-block, the partial derivative of the cost function with respect to the symbol of the kth sub-block is set to zero to obtain the symbol estimate value of the kth sub-block.

[0028] Further, the S5.3 specifically comprises: according to the cost function of the kth sub-block, the partial derivative of the cost function with respect to the channel of the kth sub-block is set to zero to obtain the estimate value of the channel tap of the kth sub-block.

[0029] Further, the S5.4 specifically comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function to the noise variance of the kth sub-block to zero to obtain the noise variance estimation value of the kth sub-block, and the result is the square of the 2-norm of the received signal residual divided by the length of the sub-block.

[0030] Further, the S5.5 comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function to the first-order Markov coefficient of the kth sub-block to zero to obtain the first-order Markov coefficient estimation value of the kth sub-block, and the result is the conjugate transpose of the k-1th sub-block channel tap estimation value multiplied by the kth sub-block channel tap estimation value, and then divided by the square of the 2-norm of the k-1th sub-block channel tap estimation value.

[0031] Further, the S5.6 specifically comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function to the channel noise variance of the kth sub-block to zero to obtain the channel noise variance estimation value of the kth sub-block, and the result is the square of the 2-norm of the kth sub-block channel residual divided by the length of the channel.

[0032] Compared with the prior art, the present application has the beneficial effects that:

[0033] 1) The present application uses a sub-block-based method, which can overcome the problems of non-cooperative arbitrary transmission, inter-channel interference and near-far effect in a distributed MIMO scene, and establish a first-order Markov model of the channel to realize reliable distributed MIMO underwater acoustic communication under a slow time-varying underwater acoustic channel.

[0034] 2) The present application derives a cost function from the perspective of maximum a posteriori probability estimation, and uses a block coordinate descent method to realize iterative estimation between symbols, channels and hyperparameters in the sub-block, so as to obtain better channel estimation and symbol detection performance in a distributed MIMO slow time-varying environment. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 is a flowchart of the present application.

[0036] Figure 2 is a schematic diagram of the real channel taps generated by simulation.

[0037] Figure 3 is the result of the 0th sub-block channel tap estimation .

[0038] Figure 4 is a flowchart of S6.

[0039] Figure 5 is a curve graph of the first-order Markov coefficient estimation value in each sub-block . ​

[0040] Figure 6 is the demodulation constellation and bit error rate under the method of the present application, wherein (a) is the demodulation constellation and bit error rate of Tx1, and (b) is the demodulation constellation and bit error rate of Tx2. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0042] As shown in the figure, a distributed MIMO single-carrier underwater acoustic communication joint channel estimation and symbol detection method comprises the following steps: Figure 1

[0043] S1: Constructing a system model: a distributed MIMO single-carrier underwater acoustic communication system is constructed, which contains M transducers and N hydrophones. The M transducers are distributed on M different transmitting nodes, each of which is independently controlled, and the transmitting data of each transmitter cannot be known by other transmitters. The N hydrophones are distributed on the same receiving end, receive signals from multiple transmitting nodes, and demodulate the signals. The speeds of the M different transmitting nodes are approximately the same, the distances from the receiving end are arbitrary, and the data transmission times of the M transducers are arbitrary. In this embodiment, M=2 and N=2. For convenience of representation, Txm and Rxn are defined as the mth transducer and the nth hydrophone, respectively. It is assumed that the data x m transmitted by Txm is:

[0044]

[0045] wherein the superscript T means transposition, represents a t m ×1-dimensional zero vector, and the parameter t m represents the time of delayed transmission of each transducer. Without loss of generality, it is assumed that t1 M , and t1=0, x m,p represents a training symbol vector transmitted by Txm, x m,d represents an information symbol vector transmitted by Txm, and the data is a zero-mean random sequence with a variance of In this embodiment, the time of delayed transmission is t1=0 and t2=1000, the length of the training symbol is L p 511, and the length of the symbol is L d 3840. The data y n received by Rxn is: ​

[0046]

[0047] Among them, w n This indicates zero mean and variance. Complex Gaussian noise, H m,n Let Txm be the channel convolution matrix between Txm and Rxn, and let h be the channel tap between Txm and Rxn. m,n Defined as:

[0048]

[0049] in, This involves constructing a diagonal matrix from the elements extracted from the matrix. In this embodiment, the simulated channel tap h... m,n As shown in Figure 2, the channel tap length L ch It is 15.

[0050] Then, the received data... Divide into sub-blocks, assuming each sub-block has a length of L. sb The overlap length between sub-blocks is L ovlp Then the received signals of the k sub-blocks of Rxn for:

[0051]

[0052] in Let Txm be the channel convolution matrix of the k-th sub-block between Txm and Rxn. The data transmitted for the k-th sub-block corresponding to Txm. This represents the zero mean and variance of the k sub-blocks received by Rxn. Complex Gaussian noise, N sb The number of sub-blocks is as follows:

[0053]

[0054]

[0055]

[0056]

[0057] in This indicates that Rxn receives the signal y. n The i-th element, Data x sent for Txm m The i-th element, This indicates that Rxn receives complex Gaussian noise w. n The i-th element, are the channel taps of Txmand Rxin the kthblock, and are the i-th elements of the channel taps of Txmand Rxin the kthsubblock, N sb is the number of subblocks. In this embodiment, the subblock length L sb is set to 40, the overlap length L ovlp between subblocks is set to 15, thus the number of subblocks N sb is 215.

[0058] S2: Constructing the channel model: To cope with the slow time-varying property of the underwater acoustic communication channel, the channel between subblocks is modeled by a first-order Markov model, i.e.,

[0059]

[0060] where are the first-order Markov coefficients of Txmand Rxin the kthblock, and are the channel taps of the kthsubblock and the (k-1)thsubblock, respectively, and is a complex Gaussian noise with zero mean and variance , which is used to model the random noise component in the channel state transition.

[0061] S3: Deriving the cost function J of the distributed MIMO single-carrier underwater acoustic communication system from the maximum a posteriori probability (MAP) perspective, based on the distributed MIMO single-carrier underwater acoustic communication system and the channel model. S3 includes the following sub-steps:

[0062] S3.1: Since there are L ovlp overlapping symbols between subblocks, part of the symbols in the current subblock can be accurately estimated in the previous subblock, thus in the current kthsubblock, the first L g symbols of the mthuser's to-be-estimated symbol can be regarded as known, and the last L u symbols are unknown, L g and L u satisfy L g + L u = L sb + L ch - 1, thus we have:

[0063]

[0064] S3.2: Reconstructing the system model of the kthsubblock as:

[0065] ​​​

[0066] wherein, is the first to the L g th column of is the L g th to the L g +L u th column of

[0067] When the symbol is known and the channel taps are unknown in the kth sub-block, the system model of the kth sub-block can be rewritten as:

[0068]

[0069] wherein, is the symbol convolution matrix of Txm in the kth sub-block, denoted as:

[0070]

[0071] wherein, v is an intermediate number, .

[0072] S3.3: According to the system model and the channel model, the maximum a posteriori probability p of k is:

[0073]

[0074] wherein e is a natural constant.

[0075] S3.4: Taking the natural logarithm and the reciprocal of the posteriori probability, the kth sub-block cost function J k is obtained:

[0076]

[0077] S4: Initialization: initializing the channel tap estimation of the 0th sub-block , the noise variance , the first-order Markov coefficient and the noise variance of the channel In the embodiment, the initializations are , , The channel tap estimation of the 0th sub-block includes the following sub-steps:

[0078] S4.1: Constructing the Toeplitz matrix X of the training symbol of the Txm transmitted data: m,p

[0079] ​

[0080] wherein, represents the i-th element of the training symbol vector X m,p .

[0081] S4.2: The Rxn received data y n corresponding to the part of the Txm training symbol vector X m,p in the k-th sub-block is:

[0082]

[0083] S4.3: The channel tap estimate of the 0-th sub-block between the m-th transducer and the n-th hydrophone can be expressed by the least square method, that is, the pseudo-inverse of the Toeplitz matrix of the training symbols of the data sent by the m-th transducer is multiplied by the part of the training symbol vector of the m-th transducer corresponding to the m-th transducer in the received data of the n-th hydrophone. The estimated channel tap of the 0-th sub-block is:

[0084]

[0085] wherein, the superscript H means conjugate transpose, and the superscript -1 means matrix inversion. In this embodiment, the channel tap estimate result of the 0-th sub-block is as shown in FIG. 3.

[0086] S5: In the k-th sub-block, first, the channel estimate, the noise variance estimate, the first-order Markov coefficient estimate, and the channel noise variance estimate are kept consistent with the k-1-th sub-block;

[0087] Then, the symbol of the k-th sub-block is estimated to obtain its estimate , and then the channel estimate, the sub-block length on the noise variance estimate, the first-order Markov coefficient estimate, and the channel noise variance estimate are updated in turn. Specifically, it includes:

[0088] S5.1: In the k-th sub-block, first, the channel estimate, the noise variance estimate, and the first-order Markov coefficient estimate are kept consistent with the k-1-th sub-block, that is, the update , , , wherein, is the channel estimate result of the k-th sub-block, is the channel tap estimate result of the k-1-th sub-block, is the first-order Markov coefficient of the Txm and the Rxn k-1-th block.

[0089] S5.2: In the k-th sub-block, based on the received signal Channel tap estimates The estimated value of noise variance variance of sign For symbols Make an estimate and obtain its estimated value. Specifically, it is based on the cost function J of the k-th sub-block. k Define variables , making J k right Taking the partial derivative and setting it to zero, we get The estimated value :

[0090]

[0091] in, , , yes The Lth g To L g +L u List, Defined as:

[0092]

[0093] Therefore, combining the known The estimated value can then be obtained. for .

[0094] S5.3: In the k-th sub-block, based on the received signal , estimated value of the symbol The estimated value of noise variance First-order Markov coefficient estimates and channel noise variance estimate Estimates of channel taps The update is performed. Specifically, it is based on the cost function J of the k-th sub-block. k ,definition , making J k right Taking the partial derivative and setting it to zero, we get The estimated value :

[0095]

[0096] in, , , The symbol estimated by the k-th sub-block Composition, represented as:

[0097]

[0098] Where v is the middle number, .

[0099] S5.4: In the k-th sub-block, based on the received signal , estimated value of the symbol Channel tap estimates and sub-block length Estimated value of noise variance The update is performed. Specifically, it is based on the cost function J of the k-th sub-block. k , making J k right Taking the partial derivative and setting it to zero, we get The estimated value :

[0100]

[0101] S5.5: In the k-th sub-block, based on the estimated value of the channel tap of the current k-th sub-block. and the estimated value of the channel tap of the (k-1)th sub-block. Estimates of the first-order Markov coefficients The update is performed. Specifically, it is based on the cost function J of the k-th sub-block. k , making J k right Taking the partial derivative and setting it to zero, we get The estimated value for:

[0102]

[0103] S5.6: In the k-th sub-block, based on the estimated value of the channel tap of the current k-th sub-block. The estimated value of the channel tap of the (k-1)th sub-block First-order Markov coefficient estimates and channel tap length L ch Channel noise variance estimate The update is performed. Specifically, it is based on the cost function J of the k-th sub-block. k , making J k right Taking the partial derivative and setting it to zero, we get The estimated value :

[0104]

[0105] S6: Iteratively repeat S5.1-S5.6 using the block coordinate descent method, as follows:Figure 4 As shown, until the iteration number reaches the maximum iteration number N iter Finally, joint channel estimation and symbol detection are realized, in this embodiment, let N iter =5, the first order Markov coefficient of each sub-block k is as shown. Figure 5

[0106] In this embodiment, the distributed MIMO underwater acoustic communication signal is demodulated, a distributed MIMO single-carrier underwater acoustic communication joint channel estimation and symbol detection method of the application is used, and the demodulation constellation diagram and the bit error rate are as shown. Figure 6 The encoding bit error rate of Tx1 is 0.23%, and the encoding bit error rate of Tx2 is 0.17%. After decoding, both can achieve zero bit error rate.

[0107] The distributed MIMO single-carrier underwater acoustic communication joint channel estimation and symbol detection method proposed by the application uses a sub-block-based algorithm, and can obtain better channel estimation and symbol detection performance in a distributed MIMO slow time-varying environment.

[0108] Although the embodiments of the application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the application, and the scope of the application is defined by the appended claims and their equivalents.​

Claims

1. A distributed MIMO single carrier underwater acoustic communication joint channel estimation and symbol detection method, characterized in that, The method comprises the following steps: S1: constructing a distributed MIMO single-carrier underwater acoustic communication system, comprising M transducers and N hydrophones; S2: constructing a channel model, the channel between sub-blocks is represented by a first-order Markov model; The first-order Markov model is represented as: ， wherein, is the first order Markov coefficient for the kth block of Txmand Rxn, and are the channel taps of the adjacent th and th sub-blocks, denotes a complex Gaussian noise with zero mean and variance , and Txmand Rxnare the mth transducer and the nth hydrophone, respectively. S3: according to the distributed MIMO single-carrier underwater acoustic communication system and the channel model, a cost function of the distributed MIMO single-carrier underwater acoustic communication system is derived from the perspective of maximum a posteriori probability; Taking natural logarithm and reciprocal of the posterior probability, the kth sub-block cost function J is obtained k is: ， In the formula, L sb L is the length of the sub-block. ch The channel tap length. For noise variance, The received signals of k sub-blocks of Rxn. The data transmitted for the k-th sub-block corresponding to Txm. These are the channel taps of Txm and Rxn in the k-th block. The variance of the sign. yes The 1st to Lth g List, yes The Lth g To L g +L u List, Let L be the channel convolution matrix of the k-th sub-block between Txm and Rxn. g This indicates the m-th user symbol to be estimated in the k-th sub-block. known symbols The number of L u This indicates the m-th user symbol to be estimated in the k-th sub-block. Symbols for the unknown part The number of; S4: initializing the estimated value of the channel tap of the 0th sub-block, the noise variance, the first-order Markov coefficient and the noise variance of the channel in the channel model; S5: in the kth sub-block, firstly, the estimated value of the channel tap, the estimated value of the noise variance, the estimated value of the first-order Markov coefficient and the estimated value of the channel noise variance are kept consistent with the (k-1) th sub-block; Then, the symbol of the kth sub-block is estimated to obtain the estimated value, and then the estimated value of the channel tap, the estimated value of the noise variance, the estimated value of the first-order Markov coefficient and the estimated value of the channel noise variance are updated in turn; According to the cost function of the kth sub-block, the partial derivative of the cost function with respect to the channel of the kth sub-block is set to zero to obtain the estimated value of the channel tap of the kth sub-block; S6: using the block coordinate descent method to iteratively repeat S5 until the maximum iteration number is reached to finally realize joint channel estimation and symbol detection.

2. The method according to claim 1, wherein, The S1 comprises that the M transducers are distributed on M different transmitting nodes, each of which is independently controlled, and each transmitter cannot know the transmission data of other transmitters; the N hydrophones are distributed on the same receiving end to receive signals from multiple transmitting nodes and demodulate; the M different transmitting nodes have a speed close to, and the distance from the receiving end is arbitrary, and the transmission data time of the M transducers is arbitrary.

3. The method according to claim 1, wherein, The initialization of the estimated value of the channel tap of the 0th sub-block in the S4 comprises: S4.1: constructing a Toeplitz matrix of the training symbol of the data transmitted by the mth transducer; S4.2: intercepting the part of the training symbol vector of the mth transducer corresponding to the data received by the nth hydrophone; S4.3: the estimated value of the channel tap of the 0th sub-block between the mth transducer and the nth hydrophone is represented by a least square method, that is, the pseudo-inverse of the Toeplitz matrix of the training symbol of the data transmitted by the mth transducer is multiplied by the part of the training symbol vector of the mth transducer corresponding to the data received by the nth hydrophone.

4. The method of claim 1, wherein, The S5 specifically comprises: S5.1: in the kth sub-block, firstly, the estimated value of the channel tap, the estimated value of the noise variance and the estimated value of the first-order Markov coefficient are kept consistent with the (k-1) th sub-block; S5.2: the symbol of the kth sub-block is estimated to obtain the estimated value, comprising: in the kth sub-block, the symbol of the kth sub-block is estimated according to the received signal, the estimated value of the channel tap, the estimated value of the noise variance and the variance of the symbol to obtain the estimated value; S5.3: updating the estimated value of the channel taps comprises: in the kth sub-block, updating the estimated value of the channel according to the received signal, the estimated value of the symbol, the estimated value of the noise variance, the estimated value of the first-order Markov coefficient and the estimated value of the channel noise variance; S5.4: updating the estimated value of the noise variance comprises: in the kth sub-block, updating the estimated value of the noise variance according to the received signal, the estimated value of the symbol, the estimated value of the channel taps and the length of the sub-block; S5.5: updating the estimated value of the first-order Markov coefficient comprises: in the kth sub-block, updating the estimated value of the first-order Markov coefficient according to the estimated value of the channel taps of the current kth sub-block and the estimated value of the channel taps of the k-1th sub-block; S5.6: updating the estimated value of the channel noise variance comprises: in the kth sub-block, updating the estimated value of the channel noise variance according to the estimated value of the channel taps of the current kth sub-block, the estimated value of the channel taps of the k-1th sub-block, the estimated value of the first-order Markov coefficient of the k-1th sub-block and the length of the channel.

5. The method according to claim 4, wherein, The S5.2 specifically comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function with respect to the symbol of the kth sub-block to zero to obtain the estimated value of the symbol of the kth sub-block.

6. The method of claim 4, wherein, The S5.4 specifically comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function with respect to the noise variance of the kth sub-block to zero to obtain the estimated value of the noise variance of the kth sub-block, and the result is the square of the 2-norm of the received signal residual divided by the length of the sub-block.

7. The method of claim 4, wherein, The S5.5 comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function with respect to the first-order Markov coefficient of the kth sub-block to zero to obtain the estimated value of the first-order Markov coefficient of the kth sub-block, and the result is the conjugate transpose of the estimated value of the channel taps of the k-1th sub-block multiplied by the estimated value of the channel taps of the kth sub-block, and then divided by the square of the 2-norm of the estimated value of the channel taps of the k-1th sub-block.

8. The method according to claim 4, wherein, The S5.6 specifically comprises: according to the cost function of the kth sub-block, setting the partial derivative of the cost function with respect to the channel noise variance of the kth sub-block to zero to obtain the estimated value of the channel noise variance of the kth sub-block, and the result is the square of the 2-norm of the channel residual of the kth sub-block divided by the length of the channel.

Images