A Hyperparameter Optimization Design Method for a Two-User Distributed MIMO Underwater Acoustic Communication Subblock

Abstract

This invention discloses a hyperparameter optimization design method for sub-blocks in a two-user distributed MIMO underwater acoustic communication system. The method includes: defining the length of data packets transmitted by the users, the staggered arrival times of the data transmitted by the two users at the receiver, and the channel length; obtaining the sub-block index that first includes user 2's data on the time axis; based on the channel length, the staggered arrival times of the data transmitted by the two users at the receiver, and the sub-block index that first includes user 2's data on the time axis, determining that within the sub-block at that index, the data lengths of both users must be greater than or equal to the channel length; defining a new index such that the last data in the sub-block of this index coincides with the last received symbol of user 1 due to multipath channel trailing. The solution is obtained using a numerical algorithm. This invention can effectively adapt to channel variations in complex marine environments and exhibits better environmental robustness in practical engineering deployments.

Description

Technical Field

[0001] This invention relates to the field of distributed MIMO underwater acoustic communication, specifically to a hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block. Background Technology

[0002] In distributed MIMO underwater acoustic communication, uncoordinated arbitrary transmission by multiple users can lead to signal collisions at the receiver. When using a sub-block-based demodulation method, without proper design of sub-block hyperparameters (i.e., sub-block length and overlap between sub-blocks), channel and data in certain sub-blocks may become unsolvable, significantly impacting the performance of multi-user data demodulation. Currently, there are no methods for designing sub-block hyperparameters in existing distributed MIMO underwater acoustic communication systems. Therefore, it is necessary to design sub-block hyperparameters for distributed MIMO underwater acoustic communication. This invention takes a two-user distributed MIMO underwater acoustic communication system as an example, analyzes the data within sub-blocks, and optimizes the hyperparameter design to improve the reliability and reduce the bit error rate of distributed MIMO underwater acoustic communication. Summary of the Invention

[0003] The purpose of this invention is to provide a hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block, so as to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block, comprising the following steps:

[0006] S1: Defines the length of the data packets sent by the user, the staggered arrival time of the data sent by the two users at the receiving end, and the channel length;

[0007] S2: Based on the staggered arrival times of the data sent by the two users at the receiving end, obtain the sub-block index on the timeline that first includes user 2's data;

[0008] S3: Based on the channel length and the staggered arrival times of the two users' transmitted data at the receiver, as well as the sub-block index that first includes user 2's data on the time axis, it is determined that within the sub-block at that index, the data length of both users must be greater than or equal to the channel length.

[0009] S4: Based on the relationship between channel length and sub-block length, the sub-block length needs to be greater than or equal to twice the channel length;

[0010] S5: Based on the relationship between the channel length and the overlap length between sub-blocks, the overlap length between sub-blocks must be greater than or equal to the channel length.

[0011] S6: Define a new index such that the last data in the sub-block of this index coincides with the last received symbol of User 1 due to the trailing effect caused by the multipath channel;

[0012] S7: Combining S2-S6, a numerical algorithm is used to solve the problem. The solution approach is a traversal algorithm.

[0013] Furthermore, S1 includes the following sub-steps:

[0014] S1.1: Length L of the data packet sent by the user blk In cooperative underwater acoustic communication systems, this is known as L. blk =L p +L d , where L p L is the length of the training symbols. d It is the length of the data symbol;

[0015] S1.2: The staggered arrival time L of the data sent by the two users at the receiving end. gap The definition is determined by the time difference between the arrival of the data sent by the user at the receiving end. The arrival time of the data sent by the user at the receiving end is calculated by the autocorrelation characteristics of the synchronization symbol.

[0016] S1.3: Channel length L ch After estimating the channel response using training symbols, the appropriate length is determined based on actual needs.

[0017] Furthermore, step S1.2 includes the following sub-steps:

[0018] S1.2.1: Assume that the data sent by user 1 arrives at the receiving end first, and the arrival time is L. arr1 Assume that the data sent by user 2 arrives at the receiving end in time L. arr2 ;

[0019] S1.2.2: Assuming the training symbols are the synchronization symbols, based on the arrival time of the data sent by the user at the receiving end, the direct path data received by the receiving end for user 1 is obtained as follows: The direct path data received by the receiver for user 2 is ,in L represents arr1 ×1-dimensional zero vector, x m,p Let x represent the training symbol vector of the m-th user. m,d This represents the data symbol vector of the m-th user. Combined with the channel, the data received by the receiver from the two users is... ,in It is a convolution symbol, h1 is the channel that user 1 experiences to reach the receiver, and h2 is the channel that user 2 experiences to reach the receiver.

[0020] S1.2.3: Based on the autocorrelation characteristics of the synchronization symbols, take the training symbol x of user 1. 1,p With the received data y str Performing cross-correlation yields the result y. ccf1 Set the threshold η and the synchronization point y. ccf1 In satisfying Symbolic index L of time syn1 `max` indicates taking the maximum value, specifically the training symbol `x` for user 2. 2,p With the received data y str Performing cross-correlation yields the result y. ccf2 Set the threshold η and the synchronization point y. ccf2 In satisfying Symbolic index L of time syn2 ;

[0021] S1.2.4: Based on the results obtained from the cross-correlation, the staggered arrival time L of the transmitted data from the two users at the receiving end is obtained. gap The expression is: L gap =L syn2 -L syn1 .

[0022] Furthermore, step S1.3 includes the following sub-steps:

[0023] S1.3.1: Construct the Topulitz matrix X of the user training symbols m,p ,have:

[0024] ,

[0025] In the formula, x m,p [n] represents the nth element of the training symbol vector of the mth user, and the preset channel length. Select based on empirical values, so that ;

[0026] S1.3.2: Estimate the channel between the m-th user and the receiver. for:

[0027] ,

[0028] In the formula, the superscript H means conjugate transpose, the superscript -1 means matrix inversion, and y p As an intermediate variable, The superscript T indicates transpose;

[0029] S1.3.3: Constructing the index vector Set a threshold δ, and sequentially index the elements of the vector. Substitution ,get ,judge and If the relationship, Then stop substituting and let q m =idx[i], then the estimated value of the channel between user m and the receiving end is idx[i]. Since m = 1, 2, let q = max{q1, q2}, therefore the updated estimates of user m and the receiver channel are... Channel length L ch For L ch =q+1.

[0030] Furthermore, S2 includes the following sub-steps:

[0031] S2.1: Define the range of received data for the k-th sub-block as follows:

[0032] ,

[0033] In the formula, L sb L represents the length of the sub-block. ovlp Indicates the overlap length between sub-blocks;

[0034] S2.2: When the k-th sub-block contains user 2's data, the last data in that sub-block... The time L between the arrival of the data sent by the two users at the receiving end gap The following relationship must be satisfied:

[0035] ;

[0036] Therefore, the range of values ​​for k is:

[0037] ;

[0038] S2.3: Define k min It is the first sub-block index on the timeline to contain user 2's data, therefore, k min The value can be:

[0039] ,

[0040] in It is an up-rounding operation.

[0041] Furthermore, S3 specifically refers to: at the k-th... min Within each sub-block, it contains user 2's data. To satisfy the requirements for estimating h2, the expression that needs to be satisfied is:

[0042] .

[0043] Furthermore, S4 specifically involves: in the sub-block-based signal processing algorithm, channel estimation is performed within each sub-block, and within the k-th sub-block, the length L of the sub-blocks it contains is... sb If L satisfies the requirements for estimating h1 and h2, then sb <2L ch If the received signal model of this sub-block is in an underdetermined state, it will lead to inaccurate channel estimation. Therefore, the expression that needs to be satisfied is:

[0044] .

[0045] Furthermore, in S5, L ovlp The following relationship must be satisfied:

[0046] .

[0047] Furthermore, S6 specifically refers to: In the sub-block-based signal processing algorithm, when a sub-block enters the tail portion of user 1, ΔL is defined as representing the data length of user 1 contained in that sub-block, i.e. When ΔL is less than the channel length L ch This will lead to inaccurate channel estimation for user 1 in that sub-block. Therefore, a constraint is added: there exists a k-th... max A sub-block, the last symbol of which is... The last symbol trailing with user 1 After overlap, channel h1 for user 1 remains unchanged until the sub-block leaves the tail of the user symbol, therefore k max The following relationship must be satisfied:

[0048] ;

[0049] k max The value can be:

[0050] ;

[0051] Based on sub-block index k max It should be an integer; the above formula can be rewritten as:

[0052] ,

[0053] in, It is a rounding operation.

[0054] Furthermore, S7 specifically refers to: specifying the range of values ​​for the sub-block length as follows: The range of values ​​for the overlap length between sub-blocks is: Using a traversal algorithm, select L that satisfies the S2-S7 relationship. sb and Lovlp The value of is the result of the sub-block hyperparameter optimization design.

[0055] Compared with the prior art, the beneficial effects of the present invention are:

[0056] 1) The present invention proposes a hyperparameter optimization design method for sub-blocks of two-user distributed MIMO underwater acoustic communication. It optimizes the design of hyperparameters such as sub-block length and overlap length between sub-blocks under two users. It can effectively adapt to channel changes in complex marine environments and can demonstrate better environmental robustness in industrial application scenarios with strong time-varying characteristics and severe multipath interference in underwater acoustic communication channels.

[0057] 2) The hyperparameter optimization design method of a two-user distributed MIMO underwater acoustic communication sub-block of the present invention establishes the mapping relationship between channel characteristics and system parameters, provides engineers with clear parameter design criteria, and solves the problem of long debugging cycle of traditional trial and error methods.

[0058] 3) The hyperparameter optimization design method of the two-user distributed MIMO underwater acoustic communication sub-block of the present invention, although designed for two-user scenarios, can be directly extended to 3-5 user scenarios, providing key technical support for building a large-scale underwater Internet of Things network and having important engineering application value.

[0059] 4) This invention provides a hyperparameter optimization design method for sub-blocks in a two-user distributed MIMO underwater acoustic communication system. Under given channel length and staggered arrival times of the two users at the receiver, the method optimizes hyperparameters such as sub-block length and overlap length between sub-blocks. This improves the reliability of distributed MIMO underwater acoustic communication and reduces the bit error rate. Attached Figure Description

[0060] Figure 1 This is a flowchart of the present invention.

[0061] Figure 2 This is a diagram illustrating the staggered arrival times of data from two households at the receiving end.

[0062] Figure 3 It represents the synchronization result and synchronization point between the two users.

[0063] Figure 4 It is the result of channel estimation. .

[0064] Figure 5 It is an estimate of the channel. .

[0065] Figure 6 This is the pseudocode of the sub-block hyperparameter optimization design method of the present invention.

[0066] Figure 7 The demodulation constellation diagram and bit error rate are after sub-block hyperparameter optimization design. (a) is the demodulation constellation diagram and bit error rate of user 1, and (b) is the demodulation constellation diagram and bit error rate of user 2.

[0067] Figure 8 The demodulation constellation diagram and bit error rate are without using sub-block hyperparameter optimization design. (a) is the demodulation constellation diagram and bit error rate of user 1, and (b) is the demodulation constellation diagram and bit error rate of user 2. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] like Figure 1 As shown in the figure, a hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block in this embodiment includes the following steps:

[0070] S1: Defines the length L of the data packets sent by the user. blk The staggered arrival time L of data sent by the two users at the receiving end gap and channel length L ch S1 includes the following sub-steps:

[0071] S1.1: Length L of the data packet sent by the user blk In cooperative underwater acoustic communication systems, this is known as L. blk =L p +L d , where L p L is the length of the training symbols. d It is the length of the data symbol; in this embodiment, L p =511, L d =3840, L blk =L p +L d =4351.

[0072] S1.2: The staggered arrival time L of the data sent by the two users at the receiving end. gap The definition is determined by the time difference between the arrival times of the data sent by the two users at the receiving end. The arrival time of the data sent by the user at the receiving end is calculated by the autocorrelation characteristics of the synchronization symbol. S1.2 includes the following sub-steps:

[0073] S1.2.1: Without loss of generality, assume that the data sent by user 1 arrives at the receiving end first, and the arrival time is L. arr1 Assume that the data sent by user 2 arrives at the receiving end in time L. arr2 ,like Figure 2 As shown, here regarding L arr1 and L arr2 It's a rough location marker, but the actual location is unknown.

[0074] S1.2.2: Assuming the training symbols are the synchronization symbols, based on the arrival time of the data sent by the user at the receiving end, the direct path data received by the receiving end for user 1 is obtained as follows: T represents transpose, and the direct path data received by the receiver for user 2 is: ,in express Zero-dimensional vector, x m,p Indicates the first The training symbol vectors of each user, x m,d Indicates the first The data symbol vectors of each user, combined with the channel, result in the data received by the receiver from two users. ,in It is a convolution symbol, h1 is the channel that user 1 experiences to reach the receiver, and h2 is the channel that user 2 experiences to reach the receiver.

[0075] S1.2.3: Based on the autocorrelation characteristics of the synchronization symbols, take the training symbol x of user 1. 1,p With the received data y str Performing cross-correlation yields the result y. ccf1 Set the threshold η and the synchronization point y. ccf1 In satisfying Symbolic index L of time syn1 `max` indicates taking the maximum value, specifically the training symbol `x` for user 2. 2,p With the received data y str Performing cross-correlation yields the result y. ccf2 Set the threshold η and the synchronization point y. ccf2 In satisfying Symbolic index L of time syn2 In this embodiment, the threshold η is set to 0.6. and Results and synchronization points such as Figure 3 As shown, the synchronization point is L. syn1 =1807,L syn2 =3982.

[0076] S1.2.4: Based on the results obtained from the cross-correlation, the staggered arrival time L of the transmitted data from the two users at the receiving end is obtained. gapThe expression is: L gap =L syn2 -L syn1 ,according to Figure 3 The results show that L gap =3982-1807=2175.

[0077] S1.3: Channel length L ch The length is determined by estimating the channel response through training symbols and then truncating it to an appropriate length according to actual needs. S1.3 includes the following sub-steps:

[0078] S1.3.1: Construct the Topulitz matrix X of the user training symbols m,p ,have:

[0079] ,

[0080] In the formula, x m,p [n] represents the nth element of the training symbol vector of the mth user, and the preset channel length. Select based on empirical values, so that In this embodiment, let .

[0081] S1.3.2: Estimate the channel between the m-th user and the receiver. for:

[0082] ,

[0083] In the formula, the superscript H means conjugate transpose. The meaning is matrix inversion, y p As an intermediate variable, The superscript T indicates transpose, such as Figure 4 The result shown is the channel estimation result. .

[0084] S1.3.3: Construct the index vector idx, Set a threshold δ, and sequentially index the elements of the vector. Substitution ,get In the index value at ,judge and If the relationship, If the substitution stops, then a new variable q is defined. m , representing the index at which the circuit stops, i.e., q m =idx[i], then the estimated value of the channel between user m and the receiving end is... for Since m = 1, 2, let q = max{q1, q2}, therefore the updated estimates of user m and the receiver channel are... Channel length L ch For L ch =q+1, in this embodiment, the threshold δ=0.1 is set, resulting in q1=3, q2=45, q=45, therefore L ch =46, such as Figure 5 The figure shows the estimated value of the channel. .

[0085] S2: Based on the staggered arrival time L of the data sent by the two users at the receiving end. gap This yields the sub-block index k on the timeline that first includes user 2's data. min The expression S2 includes the following sub-steps:

[0086] S2.1: Define the range of received data for the k-th sub-block as follows:

[0087] ,

[0088] In the formula, L sb The length of the sub-block is L. ovlp Indicates the overlap length between sub-blocks.

[0089] S2.2: When the k-th sub-block contains user 2's data, the last data in that sub-block... The time L between the arrival of the data sent by the two users at the receiving end gap The following relationship should be satisfied:

[0090] .

[0091] Therefore, the range of values ​​for k is:

[0092] .

[0093] S2.3: Define k min It is the first sub-block index on the timeline to contain user 2's data, therefore, k min The value can be:

[0094] .

[0095] S3: In sub-block-based signal processing algorithms, channel estimation is required within each sub-block. Therefore, at the k-th... min Within each sub-block, it contains user 2's data. The expression that satisfies the requirements for estimating h2 is:

[0096] .

[0097] S4: In sub-block-based signal processing algorithms, channel estimation is required within each sub-block. Therefore, in the k-th sub-block, the data L it contains... sb If L can satisfy the requirements for estimating h1 and h2, then L sb <2L ch If the received signal model of this sub-block is in an underdetermined state, it will lead to inaccurate channel estimation. Therefore, the expression that needs to be satisfied is:

[0098] ,

[0099] Among them, L sb ≤L p This is because when k=1, channel estimation requires known training symbols, thus requiring a sub-block length L. sb Less than the length L of the training sequence p .

[0100] S5: In sub-block-based signal processing algorithms, during symbol detection, the last L of the sub-block... ch Insufficient parameters can lead to performance degradation for a single symbol, therefore it is necessary to re-detect the L of the previous sub-block in the current sub-block. ch The symbol represents the overlap length L between sub-blocks. ovlp It must be greater than the channel length L ch In addition, L also needs to be satisfied. ovlp <L sb To satisfy the sliding of the sub-block, therefore L ovlp The following relationship must be satisfied:

[0101] .

[0102] S6: The length of the data sent by the user is L. blk Due to the multipath effect of the channel, the data received at the receiver has a trailing effect, and the length of the received data is L. blk +L ch -1, In the sub-block-based signal processing algorithm, when a sub-block enters the tail portion of user 1, ΔL is defined as the length of user 1 data contained in that sub-block, i.e. When ΔL is less than the channel length L ch This will lead to inaccurate channel estimation for user 1 in that sub-block. Therefore, a constraint is added: there exists a k-th... max A sub-block, the last symbol of which is... The last symbol trailing with user 1 After overlap, channel h1 for user 1 remains unchanged until the sub-block leaves the tail of the user symbol, therefore k maxThe following relationship must be satisfied:

[0103] .

[0104] Therefore, k max The value can be:

[0105] .

[0106] Based on sub-block index k max The value should be an integer; the above expression can be rewritten as:

[0107] ,

[0108] in, It is a rounding operation.

[0109] Then a feasibility analysis is performed, and ΔL is:

[0110] .

[0111] Therefore, in k max In the sub-block, △L satisfies △L≥L ch This fulfills the requirement that accurate channel estimation be performed for user 1 in this sub-block.

[0112] S7: Use numerical algorithms to solve for hyperparameters, namely the length of sub-blocks and the overlap length between sub-blocks. The range of values ​​for the sub-block length is specified as follows: The range of values ​​for the overlap length between sub-blocks is: Using a traversal algorithm, select L that satisfies the S2-S6 relationship. sb and L ovlp The value of is the result of the sub-block hyperparameter optimization design. The pseudocode of the traversal algorithm is as follows: Figure 6 As shown, and L is obtained. sb and L ovlp The first 50 values ​​are shown in Table 1.

[0113] Table 1. Results of sub-block hyperparameter optimization design (L) sb and L ovlp The first 50 values

[0114]

[0115] In this embodiment, L is selected. sb =204 and L ovlp =73, and demodulate the distributed MIMO underwater acoustic communication signals of the two users. Using the hyperparameter design method of the two-user distributed MIMO underwater acoustic communication sub-block of the present invention, the demodulation constellation diagram and bit error rate are as follows. Figure 7As shown, this method allows for arbitrary values ​​of the sub-block length and the overlap length between sub-blocks, i.e., L. sb =204 and L ovlp =74, demodulation constellation diagram and bit error rate are as follows Figure 8 As shown, by using the hyperparameter design method for a two-user distributed MIMO underwater acoustic communication sub-block proposed in this invention, the bit error rate of user 1 was reduced by 1.992%, and the bit error rate of user 2 was reduced by 0.026%.

[0116] The present invention proposes a hyperparameter design method for sub-blocks in two-user distributed MIMO underwater acoustic communication, which can optimize the design of hyperparameters such as sub-block length and overlap length between sub-blocks, thereby improving the reliability of distributed MIMO underwater acoustic communication and reducing the bit error rate.

[0117] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block, characterized in that, The method includes the following steps: S1: Defines the length of the data packets sent by the user, the staggered arrival time of the data sent by the two users at the receiving end, and the channel length; S2: Based on the staggered arrival times of the data sent by the two users at the receiving end, obtain the sub-block index on the timeline that first includes user 2's data; S3: Based on the channel length and the staggered arrival times of the two users' transmitted data at the receiver, as well as the sub-block index that first includes user 2's data on the time axis, it is determined that within the sub-block at that index, the data length of both users must be greater than or equal to the channel length. S4: Based on the relationship between channel length and sub-block length, the sub-block length needs to be greater than or equal to twice the channel length; S5: Based on the relationship between the channel length and the overlap length between sub-blocks, the overlap length between sub-blocks must be greater than or equal to the channel length. S6: Define a new index such that the last data in the sub-block of this index coincides with the last received symbol of User 1 due to the trailing effect caused by the multipath channel; S7: Combining S2-S6, a numerical algorithm is used to solve the problem. The solution approach is a traversal algorithm.

2. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 1, characterized in that, S1 includes the following sub-steps: S1.1: Length L of the data packet sent by the user blk In cooperative underwater acoustic communication systems, this is known as L. blk =L p +L d , where L p L is the length of the training symbols. d It is the length of the data symbol; S1.2: The staggered arrival time L of the data sent by the two users at the receiving end. gap The definition is determined by the time difference between the arrival of the data sent by the two users at the receiving end. The arrival time of the data sent by the user at the receiving end is calculated by the autocorrelation characteristics of the synchronization symbol. S1.3: Channel length L ch After estimating the channel response using training symbols, the appropriate length is determined based on actual needs.

3. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 2, characterized in that, S1.2 includes the following sub-steps: S1.2.1: Assume that the data sent by user 1 arrives at the receiving end first, and the arrival time is L. arr1 Assume that the data sent by user 2 arrives at the receiving end in time L. arr2 ; S1.2.2: Assuming the training symbols are the synchronization symbols, based on the arrival time of the data sent by the user at the receiving end, the direct path data received by the receiving end for user 1 is obtained as follows: The direct path data received by the receiver for user 2 is ,in L represents arr1 ×1-dimensional zero vector, x m,p Let x represent the training symbol vector of the m-th user. m,d This represents the data symbol vector of the m-th user. Combined with the channel, the data received by the receiver from the two users is... ,in It is a convolution symbol, h1 is the channel that user 1 experiences to reach the receiver, and h2 is the channel that user 2 experiences to reach the receiver. S1.2.3: Based on the autocorrelation characteristics of the synchronization symbols, take the training symbol x of user 1. 1,p With the received data y str Performing cross-correlation yields the result y. ccf1 Set the threshold η and the synchronization point y. ccf1 In satisfying Symbolic index L of time syn1 `max` indicates taking the maximum value, specifically the training symbol `x` for user 2. 2,p With the received data y str Performing cross-correlation yields the result y. ccf2 Set the threshold η and the synchronization point y. ccf2 In satisfying Symbolic index L of time syn2 ; S1.2.4: Based on the results obtained from the cross-correlation, the staggered arrival time L of the transmitted data from the two users at the receiving end is obtained. gap The expression is: L gap =L syn2 -L syn1 .

4. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 2, characterized in that, S1.3 includes the following sub-steps: S1.3.1: Construct the Topulitz matrix X of the user training symbols m,p ,have: ， In the formula, x m,p [n] represents the nth element of the training symbol vector of the mth user, and the preset channel length. Select based on empirical values, so that ; S1.3.2: Estimate the channel between the m-th user and the receiver. for: ， In the formula, the superscript H means conjugate transpose, the superscript -1 means matrix inversion, and y p As an intermediate variable, The superscript T indicates transpose; S1.3.3: Constructing the index vector Set a threshold δ, and sequentially index the elements of the vector. Substitution ,get ,judge and If the relationship, Then stop substituting and let q m If =idx[i], then the estimated value of the channel between user m and the receiving end is... Since m = 1, 2, let q = max{q1, q2}, therefore the updated estimates of user m and the receiver channel are... Channel length L ch For L ch =q+1.

5. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 3, characterized in that, The S2 step includes the following sub-steps: S2.1: Define the range of received data for the k-th sub-block as follows: ， In the formula, L sb L represents the length of the sub-block. ovlp Indicates the overlap length between sub-blocks; S2.2: When the k-th sub-block contains user 2's data, the last data in that sub-block... The time L between the arrival of the data sent by the two users at the receiving end gap The following relationship must be satisfied: ， Therefore, the range of values ​​for k is: ， S2.3: Define k min It is the first sub-block index on the timeline to contain user 2's data, therefore, k min The value can be: ， in It is an up-rounding operation.

6. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 5, characterized in that, Specifically, S3 refers to: at the k-th time... min Within each sub-block, it contains user 2's data. To satisfy the requirements for estimating h2, the expression that needs to be satisfied is: 。 7. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 6, characterized in that, Specifically, S4 refers to the following: In the sub-block-based signal processing algorithm, channel estimation is performed within each sub-block. Within the k-th sub-block, the length L of the sub-blocks it contains... sb If L satisfies the requirements for estimating h1 and h2, then sb <2L ch If the received signal model of this sub-block is in an underdetermined state, it will lead to inaccurate channel estimation. Therefore, the expression that needs to be satisfied is: 。 8. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 5, characterized in that, In the S5 mentioned above, L ovlp The following relationship needs to be satisfied. 。 9. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 5, characterized in that, Specifically, S6 refers to the following: In the sub-block-based signal processing algorithm, when a sub-block enters the tail portion of user 1, ΔL is defined as the data length of user 1 contained in that sub-block, i.e. When ΔL is less than the channel length L ch This will lead to inaccurate channel estimation for user 1 in that sub-block. Therefore, a constraint is added: there exists a k-th... max A sub-block, the last symbol of which is... The last symbol trailing with user 1 After overlap, channel h1 for user 1 remains unchanged until the sub-block leaves the tail of the user symbol, therefore k max The following relationship must be satisfied: ， k max The value can be: ， Based on sub-block index k max It should be an integer; the above formula can be rewritten as: ， in, It is a rounding operation.

10. The hyperparameter optimization design method for a two-user distributed MIMO underwater acoustic communication sub-block according to claim 5, characterized in that, Specifically, S7 specifies that the range of values ​​for the sub-block length is: The range of values ​​for the overlap length between sub-blocks is: Using a traversal algorithm, select L that satisfies the S2-S7 relationship. sb and L ovlp The value of is the result of the sub-block hyperparameter optimization design.

Images