Symbol polar search method based on cross-entropy optimization

Abstract

The present application relates to low-orbit satellite communication technical field, provide a kind of symbol polarity search method based on cross entropy optimization, comprising: initialization symbol polarity probability distribution, according to symbol period is divided into K symbols to receive signal matrix, determine the correlation result matrix of receive signal matrix and local code word.Based on current symbol polarity probability distribution, randomly generate candidate symbol sequence, calculate the fitness of each candidate symbol sequence and correlation result matrix.Select the fitness of the first second number of candidate symbol sequence to form target set.According to target set, optimize symbol polarity probability distribution by cross entropy until meeting convergence condition, output symbol polarity estimation result.According to the correlation peak position and amplitude of the highest fitness, determine code phase and carrier phase respectively.Through cross entropy, symbol polarity search can not only obtain symbol polarity information, but also estimate code phase and carrier phase.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of low-orbit satellite communication, and particularly relates to a symbol polarity search method based on cross-entropy optimization. BACKGROUND

[0002] Low-orbit satellite communication system has become an important part of information infrastructure due to its wide coverage, low transmission delay and flexible networking capability. However, the satellite-ground communication link is relatively weak and the stable communication time is extremely valuable due to its long communication distance and vulnerability to interference. Due to the high-speed movement of low-orbit satellites and the power limitation of terminal devices, a spread spectrum communication system is often used for signal transmission. However, the signal is easily affected by shielding, interference and noise during transmission, resulting in frequent interruption of the communication link. Such interruption will cause instantaneous loss of data transmission, instantaneous loss of lock of the despreading tracking module, loss of lock of the code phase tracking loop at the receiving end, and failure of carrier phase synchronization, making the subsequent signal recovery face great challenges.

[0003] After the interruption occurs, the relative movement of the satellite and the terminal causes the code phase deviation to rapidly accumulate, exceeding the maintenance range of the tracking loop, so that the code phase information is completely lost. At the same time, the carrier phase is randomized during the interruption and presents a uniform distribution characteristic. In addition, since the symbol polarity information is unknown after the interruption, the receiving end cannot directly despread and demodulate the signal. This scenario of simultaneous loss of multiple parameters makes it difficult for traditional signal reacquisition methods to effectively cope with it. At the same time, in low-orbit satellite communication, the signal power is extremely low, and the single symbol energy is often overwhelmed by noise, so it is impossible to directly recover the effective information by despreading. During the reacquisition stage after the interruption, the signal has entered the data segment, and the symbol has modulation information, so it cannot be directly coherently accumulated like the training sequence. In the traditional method, although the non-coherent detection avoids the phase synchronization problem by energy accumulation, it will introduce a square noise amplification effect, which will further deteriorate the signal-to-noise ratio. SUMMARY

[0004] The present application provides a symbol polarity search method based on cross-entropy optimization to solve the defect that part of the signal cannot be normally despread when the spread spectrum signal loses lock in the despreading tracking under low signal-to-noise ratio channel conditions in the prior art. The symbol polarity search by cross-entropy not only can obtain the symbol polarity information, but also can estimate the code phase and carrier phase. The complexity of the exhaustive algorithm is avoided, and multi-symbol energy accumulation under low signal-to-noise ratio is realized. The technical scheme provided by the present application is as follows:

[0005] In a first aspect, the present application provides a symbol polarity search method based on cross-entropy optimization, comprising:

[0006] initializing a symbol polarity probability distribution;

[0007] Acquire the received signal and local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword;

[0008] A first number of candidate symbol sequences are randomly generated based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols;

[0009] Calculate the fitness of each candidate symbol sequence with respect to the correlation result matrix;

[0010] From the first number of candidate symbol sequences, select the second number of candidate symbol sequences with fitness to form the target set; wherein the first number is greater than the second number;

[0011] Based on the target set, the symbol polarity probability distribution is optimized using cross-entropy until the convergence condition is met, and the symbol polarity estimation result is output.

[0012] The code phase is determined based on the position of the correlation peak with the highest fitness, and the carrier phase is determined based on the real and imaginary parts of the correlation peak.

[0013] Optionally, determining the correlation result matrix between the received signal matrix and the local codeword includes:

[0014] Perform a Fast Fourier Transform on the received signal matrix and the local codeword to obtain the transformed received signal matrix and the transformed local codeword, respectively.

[0015] The transformed local codeword is multiplied by each column of the transformed received signal matrix to obtain the correlation result matrix.

[0016] Optionally, the fitness is determined by the following formula:

[0017] ;

[0018] in, Indicates the first The fitness of each candidate symbol sequence; and They represent dimensions as follows: and A column vector of all 1s; Indicates the Kronecker product; Represents the correlation result matrix Transpose of; Indicates the first A sequence of candidate symbols.

[0019] Optionally, the symbol polarity probability distribution is optimized using cross-entropy based on the target set until the convergence condition is met, and the symbol polarity estimation result is output, including:

[0020] In each iteration, the symbol statistical properties of the target set are determined as the target distribution; wherein, the symbol statistical properties are the mean polarity probabilities of the symbols;

[0021] Calculate the cross-entropy between the current symbol polarity probability distribution and the target distribution;

[0022] Update the probability parameters along the cross-entropy gradient descent direction until the convergence condition is met, and output the symbol polarity estimation result.

[0023] Optionally, the convergence condition is that the number of iterations reaches a preset number or the peak value of the fitness exceeds a preset threshold.

[0024] Optionally, the received signal is represented as:

[0025] ;

[0026] in, Indicates the received signal; Represents discrete time points; Indicates the power of the received signal; Represents the original symbol waveform; Indicates symbolic index; Indicates a single symbol period; Indicates the first The sampling time of each symbol; The code offset at the moment of signal reacquisition; It represents exponential operations with the natural constant e as the base; Represents the imaginary unit; Indicates the phase of a random carrier wave; This represents additive white Gaussian noise.

[0027] Secondly, the present invention also provides a symbol polarity search device based on cross-entropy optimization, comprising the following modules:

[0028] The initialization module is used to initialize the symbol polarity probability distribution;

[0029] The matrix construction module is used to acquire the received signal and the local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword.

[0030] The sequence generation module is used to randomly generate a first number of candidate symbol sequences based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols;

[0031] The fitness calculation module is used to calculate the fitness of each candidate symbol sequence with respect to the correlation result matrix;

[0032] The sequence selection module is used to select candidate symbol sequences with fitness ranking in the top second number from the first number of candidate symbol sequences to form a target set; wherein the first number is greater than the second number;

[0033] The cross-entropy optimization module is used to optimize the symbol polarity probability distribution based on the target set using cross-entropy until the convergence condition is met, and output the symbol polarity estimation result.

[0034] The phase determination module is used to determine the code phase based on the position of the correlation peak with the highest fitness, and to determine the carrier phase based on the real and imaginary parts of the correlation peak.

[0035] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the symbolic polarity search method based on cross-entropy optimization as described in the first aspect above.

[0036] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the symbol polarity search method based on cross-entropy optimization as described in the first aspect above.

[0037] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the symbolic polarity search method based on cross-entropy optimization as described in the first aspect above.

[0038] Based on the above technical solution, the beneficial effects of the present invention compared with the prior art are as follows:

[0039] This invention provides a symbol polarity search method based on cross-entropy optimization. By initializing and iteratively updating the symbol polarity probability distribution, it avoids the high computational complexity of traditional exhaustive searches for all possible symbol combinations. It employs a probability-driven approach to gradually approximate the correct solution, significantly reducing the computational burden. After acquiring the received signal and local codewords, it segments the signal and constructs a correlation result matrix, providing a unified processing framework for subsequent symbol polarity and phase estimation. Based on the current probability distribution, it generates candidate symbol sequences and calculates their fitness. High-fitness candidate symbol sequences are selected to form the target set, and the probability distribution is updated through cross-entropy optimization. The cross-entropy optimization framework integrates symbol polarity, code phase, and carrier phase parameters into a unified search space, avoiding the high complexity and long time required for global traversal searches. It re-establishes signal connections quickly while preserving the information transmitted by the symbols. Finally, it determines the code phase and carrier phase based on the correlation peak positions and phase information, proving that this method can indeed obtain estimation results for all three types of parameters simultaneously through the symbol polarity search process. The entire process replaces the traditional exhaustive search with iterative optimization of probability distribution, avoiding the combinatorial explosion problem; at the same time, it uses correlation operations to achieve multi-symbol energy accumulation, overcoming the deficiency of insufficient single-symbol energy in low signal-to-noise ratio environments.

[0040] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0041] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating the symbolic polarity search method based on cross-entropy optimization provided by the present invention.

[0044] Figure 2 This is a schematic diagram of the cross-entropy search results provided by the present invention.

[0045] Figure 3 This is a schematic diagram of the bit error rate curve provided by the present invention.

[0046] Figure 4This is a schematic diagram of the symbol polarity search device based on cross-entropy optimization provided by the present invention.

[0047] Figure 5 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0049] The following is combined Figures 1-3 The present invention describes a symbolic polarity search method based on cross-entropy optimization.

[0050] In low-Earth orbit (LEO) satellite communication systems, signal interruptions (such as due to satellite-to-ground link obstruction, antenna attitude misalignment, or equipment transient failures) will cause the receiver's code phase tracking loop to lose lock, resulting in the loss of code phase and carrier phase information. Addressing the challenges of unknown symbol polarity and difficulty in estimating code and carrier phases in LEO satellite communication, this invention proposes a symbol polarity search method based on cross-entropy optimization to estimate the symbol polarity of spread spectrum signal sequences with unknown code and carrier phases. This aims to solve the problem of partial signal loss and inability to properly despread traditional spread spectrum signals when despreading tracking loses lock under low signal-to-noise ratio channel conditions. Symbol polarity search is performed on the lost-lock portion of the sequence, and the transmitted sequence is recovered when normal despreading is not possible. By using cross-entropy for symbol polarity search, not only can symbol polarity information be obtained, but code and carrier phases can also be estimated.

[0051] At this time, the signal with modulation sequence at the receiving end (hereinafter referred to as the received signal) is represented as:

[0052] ;

[0053] in, The received signal represents the signal received at discrete time points. The baseband signal; Indicates the power of the received signal; Represents the original symbol waveform; Indicates symbolic index; It represents the period of a single symbol, that is, the duration of a single symbol; Indicates the first The sampling time of the nth symbol, i.e., the nth symbol The start time of each symbol period; This represents the code offset at the moment of signal reacquisition, and its range is... ; Indicates at the sampling time At that time, the original symbol waveform exhibits a time-shift effect due to code offset; It represents exponential operations with the natural constant e as the base; The imaginary unit; Indicates the phase of a random carrier. This represents additive white Gaussian noise, i.e., at the sampling time... The sampled values ​​of additive white Gaussian noise.

[0054] In this scenario, signal reacquisition needs to address the following core issues: signal transmission interruption causes code phase and carrier phase deviations to exceed the tracking loop's maintenance range, necessitating a re-search. , Random phase variations in the transmitted signal can lead to coherent accumulation failure. This invention proposes a probability-driven iterative search framework that integrates symbol polarity recovery, code phase, and carrier phase estimation into a joint optimization problem.

[0055] Reference Figure 1 As shown, the specific process of this symbolic polarity search method based on cross-entropy optimization is as follows:

[0056] S110, Initialize the symbol polarity probability distribution.

[0057] In communication systems, symbol polarity refers to the binary representation of a signal (e.g., 1 and 0). Initializing the symbol polarity probability distribution involves setting an initial probability for each symbol polarity (e.g., 1 or 0). Due to a lack of prior information, it is assumed that all polarity states have equal probabilities of occurrence, i.e., a uniform distribution. This invention will... The symbol polarity probability distribution of a symbol is used This indicates that the initial value of the polarity probability of each symbol is set to... This indicates that each bit is randomly assigned a value of 0 or 1. Initializing the probability distribution provides an initial sample space for subsequent random searches, ensuring the comprehensiveness and diversity of the search process.

[0058] S120. Acquire the received signal and local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword.

[0059] Receive signal By single symbol period Divide into K symbols, each symbol containing multiple chips (e.g. (each), forming a received signal matrix. Each column of the matrix corresponds to all the code chips of a symbol. By comparing the received signal matrix with the locally pre-stored codewords one by one, the correlation between the two is calculated, forming a correlation result matrix that reflects the degree of matching.

[0060] S130. Randomly generate a first number of candidate symbol sequences based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols.

[0061] Based on the current symbol polarity probability distribution Randomly generate the first quantity (e.g.) A candidate symbol sequence of (number) is generated. Each sequence represents a possible polarity arrangement. Possible combinations of symbol polarities are explored through random sampling to provide samples for subsequent fitness calculations. The number of sequences needs to be large enough to cover potential optimal solutions, but not too large to avoid exhausting computational resources. Each candidate symbol sequence contains (number)... The symbol, the first The candidate symbol sequence is represented as follows: .

[0062] S140. Calculate the fitness of each candidate symbol sequence with the correlation result matrix.

[0063] Each candidate symbol sequence is evaluated using an assessment mechanism (such as a fitness function) to measure its degree of matching with the correlation matrix of the received signal. For example, fitness can be based on metrics such as the intensity of the correlation peak or the signal-to-noise ratio (SNR) to assess the degree of matching between the candidate symbol sequence and the received signal, quantify the quality of the candidate symbol sequence, and provide a basis for subsequent selection and updates.

[0064] For each candidate symbol sequence Calculate its correlation matrix fitness:

[0065] ;

[0066] in, Indicates the first The fitness of each candidate symbol sequence; and They represent dimensions as follows: and A column vector of all 1s; Indicates the Kronecker product; Represents the correlation result matrix Transpose of; Indicates the first A sequence of candidate symbols. This operation is equivalent to coherently accumulating each symbol to extract the total correlation energy at each chip location, while preserving the code phase resolution.

[0067] S150. From the first number of candidate symbol sequences, select the second number of candidate symbol sequences with fitness to form a target set; wherein the first number is greater than the second number.

[0068] This reflects the correlation between the candidate symbol sequence and the real signal; a larger value indicates a more accurate joint estimation of symbol polarity, code phase, and carrier phase. Based on fitness, the top two performing sequences (such as the previous ones) are selected. The candidate symbol sequences (number of which are not specified) are the primary focus of optimization. The number of retained symbols is much smaller than the initial number of candidates (i.e., the first number mentioned above). By retaining only high-fitness sequences (target set) and discarding low-quality candidates, the algorithm converges faster and the search efficiency is improved.

[0069] From all candidate symbol sequences, select the first... The sequences corresponding to each fitness level constitute the target set. For example: if 100 candidate symbol sequences are generated, the top 10 sequences based on fitness (i.e., ...) are selected. ).

[0070] S160. Optimize the symbol polarity probability distribution based on the target set using cross-entropy until the convergence condition is met, and output the symbol polarity estimation result.

[0071] Based on the statistical properties of the target set (such as the frequency of sign polarity occurrence), the sign polarity probability distribution is updated using the cross-entropy minimization method. Cross-entropy is used to measure the difference between the new distribution and the target distribution; minimizing this difference brings the new distribution closer to the optimal distribution. Steps S130-S150 are repeated until the convergence condition is met (such as reaching the maximum number of iterations), at which point the final sign polarity estimation result is output.

[0072] S170, based on the fitness level with the highest peak The code phase is determined by the position of the correlation peak, and the carrier phase is determined by the real and imaginary parts of the correlation peak.

[0073] ;

[0074] in, Indicates carrier phase, and These represent the real and imaginary parts of the relevant peak, respectively. This represents the arctangent function.

[0075] The code phase is determined based on the position of the correlation peak with the highest fitness. The code phase is used to synchronize the codewords at the transmitting and receiving ends. The carrier phase is determined based on the real and imaginary parts of the correlation peak. The carrier phase is used to demodulate the information in the received signal.

[0076] This invention provides a symbol polarity search method based on cross-entropy optimization. It proposes a multi-parameter joint estimation framework based on cross-entropy optimization, using two-dimensional parameter search and synchronous cross-entropy search to search for symbol polarity, code phase, and carrier phase parameters. When despreading tracking loses lock, dynamic re-acquisition of code phase and carrier phase is achieved through cross-entropy optimization based on the initial probability distribution. This avoids the parameter reset and full-plane search caused by signal interruption in traditional methods.

[0077] Specifically, by initializing and iteratively updating the symbol polarity probability distribution, the high computational complexity of traditional exhaustive search for all possible symbol combinations is avoided. A probability-driven approach is used to gradually approximate the correct solution, significantly reducing the computational burden. After acquiring the received signal and local codewords, the signal is segmented and a correlation result matrix is ​​constructed, providing a unified processing framework for subsequent symbol polarity and phase estimation. Candidate symbol sequences are generated based on the current probability distribution, and their fitness is calculated. High-fitness candidate symbol sequences are selected to form the target set, and the probability distribution is updated through cross-entropy optimization. The cross-entropy optimization framework integrates symbol polarity, code phase, and carrier phase parameters into a unified search space, avoiding the high complexity and long time required for global traversal search. Signal connections are re-established quickly while preserving the information transmitted by the symbols. Finally, the code phase and carrier phase are determined based on the correlation peak position and phase information, proving that this method can indeed obtain the estimation results of three types of parameters simultaneously through the symbol polarity search process. The entire process replaces traditional exhaustive search with iterative optimization of the probability distribution, avoiding the combinatorial explosion problem; simultaneously, correlation operations are used to achieve multi-symbol energy accumulation, overcoming the deficiency of insufficient single-symbol energy in low signal-to-noise ratio environments. This method combines cross-entropy optimization and random search to gradually approximate the optimal symbol polarity probability distribution, thereby improving the accuracy of symbol polarity estimation and the robustness of signal demodulation.

[0078] In scenarios where signal tracking is lost or some symbols are missing, this invention uses cross-entropy-based probabilistic optimization to quickly and effectively search for symbol polarity, avoiding system resets and information loss caused by interruptions in traditional methods. Through target set filtering and probabilistic iterative updates, the complexity of traditional three-dimensional parameter search is reduced from exponential to linear-logarithmic levels, meeting the low complexity and real-time requirements of low-Earth orbit satellite terminals.

[0079] Optionally, performing correlation operations directly in the time domain has high computational complexity, especially when the received signal matrix and codeword length are large, making real-time performance difficult to guarantee. Therefore, this invention uses Fast Fourier Transform (FFT) to convert time-domain operations into frequency-domain operations, thereby significantly reducing computational complexity. The determination of the correlation result matrix between the received signal matrix and the local codeword in S120 above includes:

[0080] Perform a Fast Fourier Transform on the received signal matrix and the local codeword to obtain the transformed received signal matrix and the transformed local codeword, respectively. Perform a dot product operation between the transformed local codeword and each column of the transformed received signal matrix to obtain the correlation result matrix.

[0081] The above received signal matrix The dimension is , For chip length, The number of symbols. The dimension of the local codeword C is... , represents the spreading code of a symbol.

[0082] For the received signal matrix Perform an FFT on each row to obtain the transformed received signal matrix RFFT. Perform an FFT on the local codeword C to obtain the transformed local codeword CFFT.

[0083] The transformed local codeword CFFT is multiplied by each column of the transformed received signal matrix RFFT. The multiplication result is then transformed back to the time domain using an inverse fast Fourier transform (IFFT) to obtain the correlation matrix. .

[0084] The time-domain computation complexity is The complexity of frequency domain correlation operations is This significantly reduces computational load. The parallel computing capability of FFT makes hardware implementation more efficient, suitable for high-speed communication scenarios. Frequency domain operations can suppress some noise through filtering characteristics, improving the signal-to-noise ratio of the correlation results. It supports multi-symbol parallel processing, adapting to different spreading factors and modulation schemes.

[0085] Optionally, the step S160 above, which optimizes the symbol polarity probability distribution based on the target set using cross-entropy until the convergence condition is met, and outputs the symbol polarity estimation result, includes:

[0086] S1601. In each iteration, the symbol statistical characteristics of the target set are determined as the target distribution; wherein, the symbol statistical characteristics are the mean polarity probabilities of the symbols.

[0087] In each iteration, the algorithm performs the following operations to determine the target distribution:

[0088] (1) Statistical analysis of the symbol polarity of all candidate symbol sequences in the target set;

[0089] (2) Calculate the frequency of 1 appearing on each symbol to obtain the mean polarity probability of the symbol. ;

[0090] ;

[0091] in, Indicates the first The mean polarity probability of each symbol; This represents the second quantity, i.e., the number of candidate symbol sequences in the target set; Indicates the first in the target set A sequence of candidate symbols; Represents the target set; Indicates the first in the target set candidate symbol sequences The Middle The values ​​that a symbol can take; It indicates that it belongs to.

[0092] (3) Average the polarity probabilities of each symbol , which serves as the target probability value for the symbol;

[0093] In this way, the symbolic statistical properties of the target set are quantified into a probability distribution, reflecting the spatial characteristics of the current optimal solution.

[0094] S1602. Calculate the cross-entropy between the current symbol polarity probability distribution and the target distribution; update the probability parameters along the cross-entropy gradient descent direction until the convergence condition is met, and output the symbol polarity estimation result.

[0095] The current symbol polarity probability distribution, i.e. The target distribution is as described above. The comparison shows that the optimization objective is to minimize the difference between the two distributions, i.e., to minimize the cross-entropy. :

[0096] ;

[0097] Calculate the gradient of the cross-entropy with respect to the probability parameters, and adjust the probability parameters along the gradient descent direction, as described above. This optimization process ensures that the probability distribution evolves in a direction more consistent with the target distribution.

[0098] The probability parameters are updated in the following way:

[0099] ;

[0100] in, Indicates the first The polarity probability of each symbol; Indicates the first The mean polarity probability of each symbol; ; This represents the learning rate.

[0101] pass The coefficients retain some information about the current probability distribution, through The coefficients introduce new information about the target distribution, and a smooth transition is achieved through weighted averaging. This update mechanism both incorporates new optimization information and maintains the stability of the algorithm.

[0102] The algorithm determines convergence based on the following criteria: the sign polarity probability approaches the boundary value of 0 or 1, the peak fitness reaches a preset threshold, or the maximum number of iterations is reached. The iteration terminates when any one of these conditions is met, and the final sign polarity estimation result is output.

[0103] This invention utilizes a cross-entropy optimization framework to simultaneously address the estimation of symbol polarity, code phase, and carrier phase, avoiding the error accumulation caused by the step-by-step processing of traditional methods. Compared to exhaustive search methods, this approach reduces computational complexity from exponential to linear-logarithmic levels, significantly improving computational efficiency. Iterative optimization of probability distributions and multi-symbol energy accumulation maintain reliable performance even under low signal-to-noise ratio conditions. Through probability updates along the gradient descent direction, the algorithm converges to a satisfactory solution in fewer iterations. The weighted update mechanism of probability parameters ensures the stability of the algorithm during optimization, preventing it from getting trapped in local optima.

[0104] Figure 2 This is a schematic diagram of the cross-entropy search results provided by this invention. The following is in conjunction with… Figure 2 The process of the symbolic polarity search method based on cross-entropy optimization is explained as follows:

[0105] 1. Initialize the symbol polarity probability distribution to a uniform distribution: symbol polarity probability distribution vector = [0.5, ...,0.5], where the polarity probability of each symbol is initially set to 0.5.

[0106] 2. Sequence Random Generation: The received signal is divided into K symbols according to the symbol period, and each symbol contains multiple chips to obtain the received signal matrix. The correlation matrix between the received signal matrix and the local codeword is then determined. Randomly generated based on the current symbol polarity probability distribution. Group candidate symbol sequence ( Figure 2 Each row in the table on the right represents a candidate symbol sequence.

[0107] 3. Coherent Accumulation: For each candidate symbol sequence, calculate the correlation matrix between the candidate symbol sequence and the correlation result matrix. The fitness is obtained by coherently accumulating each symbol and extracting the total correlation energy at each chip location. If the peak fitness reaches the preset threshold V T If the iteration stops, then stop.

[0108] 4. Filter the target set: From Among the candidate symbol sequences, the preferred fitness is ranked first. The candidate symbol sequence of the group yields the target set.

[0109] 5. Calculation: Statistically analyze the symbol polarity of all candidate symbol sequences in the target set, calculate the frequency of 1 occurrences for each symbol, and obtain the mean polarity probability of the symbols. .

[0110] 6. Iterative Loop: In each iteration, the average polarity probability of each symbol is calculated. The target probability value of this symbol is used to obtain the target distribution. The current symbol polarity probability distribution ( ,..., , Compare with the target distribution, according to Update the probabilities to obtain a new generation of probability distributions. ,..., , Repeat steps 2-5 in the iterative loop until the convergence condition is met, and output the sign polarity estimation result.

[0111] in, subscript Indicates the first A sequence of candidate symbols, index Indicates the first in the sequence A symbol, , The sequence length is given. Indicates the first During the nth iteration, the 1st The first candidate symbol sequence Probability estimates of the polarity of each symbol. Superscript This represents the next-generation probability distribution obtained through the current iteration. Indicates the first During the nth iteration, the 1st The first candidate symbol sequence Probability estimates of the polarity of each symbol. Subscript Indicates the carrier phase associated with the candidate symbol sequence. Indicates the first During the nth iteration, the 1st The probability distribution parameters of the carrier phase correction term for each candidate symbol sequence. Indicates the first During the nth iteration, the 1st The probability distribution parameters of the carrier phase correction term for each candidate symbol sequence. The subscript of indicates the first in the target distribution. Each component (corresponding to a symbol or phase).

[0112] Reference Figure 3 The figure shows the bit error rate curves for different K values. The horizontal axis represents the signal-to-noise ratio range. (dB) Energy per bit The vertical axis represents the noise power spectral density, and the horizontal axis represents the bit error rate (BER). Effective communication is maintained even at a signal-to-noise ratio as low as -4dB (when K=64). The measured bit error rate (BER) at a signal-to-noise ratio (SNR) of 6 dB differs from the theoretical BER of the BPSK system by less than 0.5 dB, demonstrating that the parameter estimation efficiency of the cross-entropy optimization framework is close to that of an ideal receiver. The BER of the sequence obtained through the cross-entropy search of this invention is essentially consistent with the theoretical BER of the BPSK system, indicating that the method of this invention can achieve efficient searching for symbol polarity.

[0113] The symbol polarity search device based on cross-entropy optimization provided by the present invention will be described below. The symbol polarity search device based on cross-entropy optimization described below can be referred to in correspondence with the symbol polarity search method based on cross-entropy optimization described above.

[0114] The symbol polarity search device based on cross-entropy optimization provided by this invention refers to... Figure 4 As shown, it includes:

[0115] Initialization module 210 is used to initialize the symbol polarity probability distribution;

[0116] The matrix construction module 220 is used to acquire the received signal and the local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword.

[0117] The sequence generation module 230 is used to randomly generate a first number of candidate symbol sequences based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols;

[0118] Fitness calculation module 240 is used to calculate the fitness of each candidate symbol sequence with respect to the correlation result matrix;

[0119] The sequence selection module 250 is used to select candidate symbol sequences with fitness ranking in the top second number from the first number of candidate symbol sequences to form a target set; wherein the first number is greater than the second number;

[0120] The cross-entropy optimization module 260 is used to optimize the symbol polarity probability distribution based on the target set through cross-entropy until the convergence condition is met, and output the symbol polarity estimation result.

[0121] The phase determination module 270 is used to determine the code phase based on the position of the correlation peak with the highest fitness, and to determine the carrier phase based on the real and imaginary parts of the correlation peak.

[0122] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340. The processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions from the memory 330 to execute a symbolic polarity search method based on cross-entropy optimization.

[0123] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0124] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program that can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to execute the symbolic polarity search method based on cross-entropy optimization provided by the above methods.

[0125] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the symbolic polarity search method based on cross-entropy optimization provided by the methods described above.

[0126] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A symbolic polarity search method based on cross-entropy optimization, characterized in that, include: Initialize the symbol polarity probability distribution; Acquire the received signal and local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword; A first number of candidate symbol sequences are randomly generated based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols; Calculate the fitness of each candidate symbol sequence with respect to the correlation result matrix; From the first number of candidate symbol sequences, select the second number of candidate symbol sequences with fitness to form the target set; wherein the first number is greater than the second number; Based on the target set, the symbol polarity probability distribution is optimized using cross-entropy until the convergence condition is met, and the symbol polarity estimation result is output. The code phase is determined based on the position of the correlation peak with the highest fitness, and the carrier phase is determined based on the real and imaginary parts of the correlation peak.

2. The symbolic polarity search method based on cross-entropy optimization according to claim 1, characterized in that, Determining the correlation result matrix between the received signal matrix and the local codeword includes: Perform a Fast Fourier Transform on the received signal matrix and the local codeword to obtain the transformed received signal matrix and the transformed local codeword, respectively. The transformed local codeword is multiplied by each column of the transformed received signal matrix to obtain the correlation result matrix.

3. The symbolic polarity search method based on cross-entropy optimization according to claim 1, characterized in that, The fitness is determined by the following formula: ； in, Indicates the first The fitness of each candidate symbol sequence; and They represent dimensions as follows: and A column vector of all 1s; Indicates the Kronecker product; Represents the correlation result matrix Transpose of; Indicates the first A sequence of candidate symbols.

4. The symbolic polarity search method based on cross-entropy optimization according to claim 1, characterized in that, Based on the target set, the symbol polarity probability distribution is optimized using cross-entropy until the convergence condition is met, and the symbol polarity estimation results are output, including: In each iteration, the symbol statistical properties of the target set are determined as the target distribution; wherein, the symbol statistical properties are the mean polarity probabilities of the symbols; Calculate the cross-entropy between the current symbol polarity probability distribution and the target distribution; Update the probability parameters along the cross-entropy gradient descent direction until the convergence condition is met, and output the symbol polarity estimation result.

5. The symbolic polarity search method based on cross-entropy optimization according to claim 1 or 4, characterized in that, The convergence condition is that the number of iterations reaches a preset number or the peak value of the fitness exceeds a preset threshold.

6. The symbolic polarity search method based on cross-entropy optimization according to claim 1, characterized in that, The received signal is represented as: ； in, Indicates the received signal; Represents discrete time points; Indicates the power of the received signal; Represents the original symbol waveform; Indicates symbolic index; Indicates a single symbol period; Indicates the first The sampling time of each symbol; The code offset at the moment of signal reacquisition; It represents exponential operations with the natural constant e as the base; Represents the imaginary unit; Indicates the phase of a random carrier wave; This represents additive white Gaussian noise.

7. A symbolic polarity search device based on cross-entropy optimization, characterized in that, include: The initialization module is used to initialize the symbol polarity probability distribution; The matrix construction module is used to acquire the received signal and the local codeword, divide the received signal into K symbols according to the symbol period, each symbol contains multiple code chips, obtain the received signal matrix, and determine the correlation result matrix between the received signal matrix and the local codeword. The sequence generation module is used to randomly generate a first number of candidate symbol sequences based on the current symbol polarity probability distribution; wherein, the candidate symbol sequence contains K symbols; The fitness calculation module is used to calculate the fitness of each candidate symbol sequence with respect to the correlation result matrix; The sequence selection module is used to select candidate symbol sequences with fitness ranking in the top second number from the first number of candidate symbol sequences to form a target set; wherein the first number is greater than the second number; The cross-entropy optimization module is used to optimize the symbol polarity probability distribution based on the target set using cross-entropy until the convergence condition is met, and output the symbol polarity estimation result. The phase determination module is used to determine the code phase based on the position of the correlation peak with the highest fitness, and to determine the carrier phase based on the real and imaginary parts of the correlation peak.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the symbolic polarity search method based on cross-entropy optimization as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the symbolic polarity search method based on cross-entropy optimization as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the symbolic polarity search method based on cross-entropy optimization as described in any one of claims 1 to 6.

Images