Detection and decoding joint iteration method and device

Abstract

The invention provides a detection and decoding joint iteration method and device, and the method comprises the steps: updating symbol-level detection information based on channel characteristics under the detection of internal loop iteration; under decoding inner loop iteration, bit-level decoding information is updated based on the code word constraint of the target symbol; through outer loop iteration, information updated in the detection inner loop and information updated in the decoding inner loop are interacted until the number of times of outer loop iteration reaches the maximum number of times of iteration, and a detection decoding result is determined based on symbol-level update detection information and bit-level update decoding information corresponding to the maximum number of times of iteration; wherein the detection inner loop iteration and the decoding inner loop iteration are executed in parallel. According to the method, complicated exponential operation and logarithm operation are not needed, the calculation complexity is effectively reduced, and calculation resources are saved. In addition, detection decoding internal loop parallel execution can accelerate result convergence and reduce time complexity, thereby remarkably improving the overall efficiency of detection decoding joint iteration.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and particularly to an iterative method and device for joint detection and decoding. Background Art

[0002] Iterative Detection and Decoding (IDD) are two key steps for a receiving end to process signals in a communication system. Among them, detection generally refers to recovering the transmitted symbol or bit sequence from the received signal, and decoding refers to further decoding these recovered symbol or bit sequences to correct possible errors and finally obtain the original transmitted information.

[0003] Existing detection and decoding schemes are mainly divided into two categories: linear detection IDD algorithms and non-linear detection IDD algorithms. Linear detection IDD algorithms face the problem of high computational complexity, and the EP-IDD algorithm of non-linear detection also has the problems of high iterative operation complexity and processing delay. Summary of the Invention

[0004] The present invention provides an iterative method and device for joint detection and decoding to solve the defects existing in the prior art.

[0005] The present invention provides an iterative method for joint detection and decoding, including the following steps: Under the inner-loop iteration of detection, update the symbol-level detection information based on the channel characteristics; Under the inner-loop iteration of decoding, update the bit-level decoding information based on the codeword constraint of the target symbol; Through the outer-loop iteration, interact the information updated respectively in the inner-loop of detection and the inner-loop of decoding until the number of outer-loop iterations reaches the maximum number of iterations, and then determine the detection and decoding result based on the symbol-level updated detection information and the bit-level updated decoding information corresponding to the maximum number of iterations; Wherein, the inner-loop iteration of detection and the inner-loop iteration of decoding are executed in parallel.

[0006] According to the iterative method for joint detection and decoding provided by the present invention, the step of updating the symbol-level detection information based on the channel characteristics under the inner-loop iteration of detection includes: Under the current inner-loop iteration of detection, determine the posterior mean of the transmitting antenna corresponding to the target symbol, and the posterior mean refers to the estimated expected value of the target symbol; the initial value of the posterior mean is determined based on the noise variance of the received signal and the channel characteristics; Update the detection mean based on the posterior mean and the channel characteristics; Update the symbol-level detection information based on the updated detection mean and the channel characteristics; Convert the updated symbol-level detection information into bit-level detection information.

[0007] According to a joint iterative method of detection and decoding provided by the present invention, the posterior mean is updated based on the following steps: Based on the symbol-level detection information and the symbol-level decoding information, update the posterior information of the target symbol in the current detection inner-loop iteration. The posterior information represents the probability or likelihood ratio that the target symbol is correctly decoded, and the initial value of the posterior information is extracted from the received signal. Based on the updated posterior information and the codeword constraint, update the posterior mean.

[0008] According to a joint iterative method of detection and decoding provided by the present invention, the conversion of the updated symbol-level detection information into bit-level detection information includes: Based on the bit values at each position of each symbol in the constellation point set, accumulate the symbol information with the same bit value at the same position, and convert it to the bit-level detection information corresponding to the position.

[0009] According to a joint iterative method of detection and decoding provided by the present invention, the interaction of the information updated respectively in the detection inner-loop and the decoding inner-loop through the outer-loop iteration includes: Through the decoding inner-loop iteration, based on the bit-level detection information, update the bit-level decoding information, and convert the bit-level decoding information into symbol-level decoding information. Through the outer-loop iteration, based on the difference between the bit-level detection information and the bit-level decoding information, update the symbol-level decoding information.

[0010] According to a joint iterative method of detection and decoding provided by the present invention, it further includes: After the interaction of the information updated respectively in the detection inner-loop and the decoding inner-loop through the outer-loop iteration, both the detection inner-loop iteration and the decoding inner-loop iteration use the interaction information as the prior information to assist the information iteration in their respective inner-loops.

[0011] The present invention also provides a joint iterative device for detection and decoding, including the following modules: A detection inner-loop module, configured to update symbol-level detection information based on channel characteristics in the detection inner-loop iteration. A decoding inner-loop module, configured to update bit-level decoding information based on the codeword constraint of the target symbol in the decoding inner-loop iteration. An outer loop module is used to interact, through outer loop iteration, the information updated respectively in the detection inner loop and the decoding inner loop until the number of outer loop iterations reaches the maximum number of iterations. Then, based on the symbol-level updated detection information and the bit-level updated decoding information corresponding to the maximum number of iterations, the detection and decoding result is determined. Wherein, the detection inner loop iteration and the decoding inner loop iteration are executed in parallel.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the detection and decoding joint iteration method as described in any one of the above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the detection and decoding joint iteration method as described in any one of the above is implemented.

[0014] The present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the detection and decoding joint iteration method as described in any one of the above is implemented.

[0015] In the detection and decoding joint iteration method and device provided by the present invention, the detection inner loop iteration updates the symbol-level detection information based on the channel characteristics, and the decoding inner loop iteration updates the bit-level decoding information based on the codeword constraint of the target symbol. That is, the detection inner loop iteration and the decoding inner loop iteration do not need to rely on the outer loop to update the corresponding information. That is, the detection inner loop iteration and the decoding inner loop iteration can be executed in parallel, and the update iterations are carried out simultaneously to jointly update the corresponding information, thereby significantly reducing the overall processing time, accelerating the convergence of the iteration process, and improving the detection and decoding efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 It is a flowchart of the detection and decoding joint iteration method provided by the present invention.

[0018] Figure 2 It is a schematic diagram of the traditional iteration model and the parallel iteration model of the present invention.

[0019] Figure 3 It is a flowchart of another detection and decoding joint iteration method provided by the present invention.

[0020] Figure 4 It is a schematic structural diagram of the detection and decoding joint iteration device provided by the present invention.

[0021] Figure 5 It is a schematic structural diagram of the electronic device provided by the present invention. Detailed implementation manners

[0022] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0023] Currently, detection and decoding schemes are mainly divided into two categories: linear detection IDD algorithms and nonlinear detection IDD algorithms. In linear detection IDD algorithms, the minimum mean square error parallel interference cancellation (MMSE-PIC) algorithm has the problem of high computational complexity. The EP-IDD algorithm for nonlinear detection has the problems of high iterative operation complexity and processing delay.

[0024] In addition, block expectation propagation (BEP) is also applied to the EP-IDD system. This method calculates the mean and variance through the soft messages fed back by the decoder to update the prior information in detection. However, this improved algorithm has the problem of performance loss. The double expectation propagation (DEP) scheme considers moment matching in both loops to obtain a better posterior estimate to improve performance, but the moment matching in each loop involves the inversion of large-dimensional matrices, resulting in too high computational complexity. To reduce the excessive complexity brought by posterior update in each loop, some IDD schemes adopt expectation propagation approximation (EPA) detection, called EPA-IDD, to reduce the complexity in the detection loop. However, all the above-mentioned EP-IDD strategies involved require a large number of exponential multiplication operations when implemented. In addition, the repeated matrix inversion for posterior estimation in each loop also leads to an increase in computational complexity.

[0025] In view of this, the present invention provides a detection and decoding joint iteration method, aiming to eliminate the inversion operation of large-dimensional matrices in the two loops of detection and decoding in the algorithm, thereby greatly reducing the algorithm complexity, and further improving the detection and decoding efficiency. In addition, the parallel execution of the inner loop of detection and decoding can accelerate the structural convergence and reduce the time complexity, thereby significantly improving the overall efficiency of detection and decoding joint iteration. Among them, Figure 1 It is a schematic flow diagram of the detection and decoding joint iteration method provided by the present invention. As Figure 1 shown, this method includes step 110, step 120 and step 130.

[0026] Step 110: Under the detection inner-loop iteration, update the symbol-level detection information based on the channel characteristics.

[0027] Here, the detection inner-loop iteration refers to, during the signal processing, making multiple detection attempts on the received signal. Each iteration improves the detection accuracy based on the previous detection results and the channel characteristics. The detection inner-loop iteration focuses on recovering the transmitted symbols as accurately as possible from the received signal. The channel characteristics describe how the signal propagates through the physical medium (such as air, optical fiber, etc.) during transmission and how these propagation conditions affect the signal strength and phase. The channel characteristics usually include noise, attenuation, phase distortion, etc. The symbol-level detection information refers to the estimation of each symbol (usually bits or groups of bits in digital communication) based on the received signal and the channel characteristics during the detection process. The symbol-level detection information includes the reliability measure of the symbol (such as probability or likelihood ratio).

[0028] Under the detection inner-loop iteration, each iteration improves the detection accuracy based on the previous detection results (such as the symbol-level detection information and bit-level detection information under the previous detection inner-loop iteration) and the channel characteristics. Through multiple iterations, the true transmitted symbols can be gradually approximated, thereby improving the detection accuracy.

[0029] Step 120: Under the decoding inner-loop iteration, update the bit-level decoding information based on the codeword constraints of the target symbol.

[0030] Specifically, the decoding inner-loop iteration means that after the symbols are detected, the decoding process tries to recover the original information according to these symbols and coding rules (such as convolutional codes, LDPC codes, etc.). The decoding inner-loop iteration refers to making multiple attempts at this process, and each iteration improves the decoding accuracy based on the previous decoding results and the reliability information of the symbols.

[0031] In addition, the target symbol refers to the symbol that is currently being attempted to be decoded. The codeword constraints refer to the rules introduced during the encoding process to ensure that the transmitted symbol sequence meets specific structural requirements (such as periodicity, redundancy, etc.). These constraints are used in the decoding process to help distinguish possible symbol sequences, thereby improving the decoding accuracy. Among them, in a communication system, there is usually a set of predefined codewords, and these codewords form the set of legal transmitted symbols. Based on the target symbol and the codeword set, a series of constraint conditions are established, and these constraint conditions are used to ensure that the decoded symbol sequence meets specific coding rules and is as close as possible to the original transmitted symbol sequence. The bit-level decoding information refers to the estimation and reliability measure of each bit (or group of bits) during the decoding process, and the bit-level decoding information is usually used to determine the final decoding result.

[0032] After applying the codeword constraint, bit-level information is extracted from the corrected symbol sequence. According to the extracted bit-level information, the decoding information is updated, and the updated decoding information will be used as the input for the next iteration of detection to further improve the accuracy of symbol-level detection and bit-level decoding. Through multiple iterations, the true transmitted bit sequence can be gradually approximated.

[0033] Step 130: Through the outer loop iteration, the information updated respectively in the detection inner loop and the decoding inner loop is interacted until the number of outer loop iterations reaches the maximum number of iterations. Then, based on the symbol-level updated detection information and bit-level updated decoding information corresponding to the maximum number of iterations, the detection and decoding result is determined.

[0034] In each outer loop iteration, the detector uses the bit-level decoding information provided by the decoder to improve the symbol-level detection information, while the decoder uses the symbol-level detection information provided by the detector to improve the bit-level decoding information. This process is repeated until the predetermined maximum number of iterations is reached.

[0035] After reaching the maximum number of iterations, based on the symbol-level detection information and bit-level decoding information obtained in the last iteration, the final detection and decoding result is determined. This result is usually the original data recovered from the received signal.

[0036] Among them, the above-mentioned detection inner loop iteration and decoding inner loop iteration are executed in parallel. Compared with the sequential execution of the detection inner loop and the decoding inner loop in the traditional method, the embodiments of the present invention can significantly reduce the overall processing time and accelerate the convergence of the iteration process.

[0037] For the joint iterative method of detection and decoding provided by the embodiments of the present invention, the detection inner loop iteration updates the symbol-level detection information based on the channel characteristics, and the decoding inner loop iteration updates the bit-level decoding information based on the codeword constraint of the target symbol. That is, the detection inner loop iteration and the decoding inner loop iteration do not need to rely on the outer loop to update the corresponding information, that is, the detection inner loop iteration and the decoding inner loop iteration can be executed in parallel, and the update iterations are carried out simultaneously to jointly update the corresponding information, so as to significantly reduce the overall processing time, accelerate the convergence of the iteration process, and improve the detection and decoding efficiency.

[0038] Based on the above embodiments, in the detection inner loop iteration, updating the symbol-level detection information based on the channel characteristics includes: In the current detection inner loop iteration, determine the posterior mean of the transmit antenna corresponding to the target symbol. The posterior mean refers to the estimated expected value of the target symbol; the initial value of the posterior mean is determined based on the noise variance of the received signal and the channel characteristics; Based on the posterior mean and the channel characteristics, update the detection mean; Based on the updated detection mean and the channel characteristics, update the symbol-level detection information; Convert the updated symbol-level detection information into bit-level detection information.

[0039] Specifically, the posterior mean refers to the estimated value of the transmitted symbol under the condition of known observation data (i.e., the received signal) and channel characteristics. The detection mean refers to the expected value of the estimated signal detection result under the condition of a given series of channel received information, posterior mean, noise variance, and channel matrix. The detection mean is used to reflect the optimal estimate of the transmitted signal under the current conditions.

[0040] Using the posterior mean and channel characteristics, a more accurate detection mean can be calculated. Using the updated detection mean and channel characteristics, the symbol-level detection information can be further updated.

[0041] In addition, the bit-level detection information refers to the detection result and its reliability metric obtained by converting the symbol-level detection result into the bit-level. The bit-level detection information will be used in the subsequent decoding process to recover the transmitted bit sequence.

[0042] Among them, the detection mean can be updated based on the following formula: Among them, represents the detection mean of the th iteration. The diag(X) function represents extracting the diagonal elements of the matrix X. represents the value of the th row and th column of the channel autocorrelation matrix, represents the diagonal correlation coefficient of the th received antenna channel autocorrelation matrix, represents the sum of the received symbols of the th received antenna, represents the number of transmit antennas.

[0043] In the above formula represents the posterior mean of the th transmit antenna in the th iteration, where the initialization

[0044] Among them, represents the initial value of the posterior mean of the th transmit antenna, , , , represents the channel matrix, represents the noise variance, represents the received signal.

[0045] The symbol-level detection information can be updated based on the following formula: where represents the symbol-level detection information of the target symbol at the -th iteration, , represents the channel matrix, represents the noise variance, represents the target symbol, represents the detection mean at the -th iteration.

[0046] Based on any of the above embodiments, the posterior mean is updated based on the following steps: Based on the symbol-level detection information and the symbol-level decoding information, update the posterior information of the target symbol in the current detection inner-loop iteration. The posterior information represents the probability or likelihood ratio that the target symbol is correctly decoded, and the initial value of the posterior information is extracted from the received signal; Based on the updated posterior information and the codeword constraint, update the posterior mean.

[0047] Considering the execution process of the EP-IDD algorithm, whether it is the internal detection iteration link or the external decoding iteration link, it involves the inversion operation of large-scale matrices. Such operations are extremely time-consuming and significantly increase the complexity of the entire algorithm. In the embodiments of the present invention, for the internal and external iteration loops in the EP-IDD algorithm, the update method of the posterior information is re-conceived and designed. Through this innovative re-design, the method successfully eliminates the originally necessary matrix inversion step, and during the implementation of the update formula, a large number of exponential operations and multiplication operations are further avoided. Thus, the operation complexity of the entire algorithm is significantly reduced.

[0048] Specifically, the posterior information of the target symbol represents the probability or likelihood ratio that the target symbol is correctly decoded. The symbol-level detection information is extracted from the received signal, which reflects the characteristics of the signal, such as amplitude, phase, frequency, etc. The symbol-level decoding information is a preliminary estimate of the transmitted bit sequence based on the known coding rules and the received symbol-level detection information.

[0049] Since the symbol-level detection information and the symbol-level decoding information are two different information sources, they each contain different aspects of information about the transmitted symbol or bit. By fusing these information, a more comprehensive and accurate posterior estimate of the transmitted sequence can be obtained, that is, based on the symbol-level detection information and the symbol-level decoding information, the posterior information can be updated.

[0050] In addition, a codeword constraint refers to a set of legal symbol sequences or bit sequences determined by coding rules. The updated posterior information contains the latest estimate of the transmitted symbol or bit and its reliability metric. Combining the updated posterior information with the codeword constraint can further constrain and correct the estimated value of the posterior mean, making it closer to the true transmitted sequence.

[0051] Among them, the posterior information can be updated with reference to the following formula: Among them, represents the posterior information of the target symbol at the -th iteration, represents the detection information of the target symbol at the -th iteration, represents the decoding information of the target symbol at the -th iteration.

[0052] It can be seen that the embodiment of the present invention updates the posterior information of the target symbol based on the symbol-level detection information and symbol-level decoding information of the target symbol, eliminating the complex matrix inversion operation.

[0053] After updating the posterior information, the posterior mean is updated based on the updated posterior information and the codeword constraint. Among them, the posterior mean can be updated with reference to the following formula: Among them, represents the posterior mean of the -th iteration, represents the codeword constraint of the target symbol , which can be understood as the set of symbols to which the target symbol belongs, represents the target symbol, represents the -th constellation point symbol of the modulation scheme.

[0054] Based on any of the above embodiments, converting the updated symbol-level detection information into bit-level detection information includes: Based on the bit values of each position of each symbol in the constellation point set, the symbol information with the same bit value at the same position is accumulated, and the corresponding bit-level detection information is obtained through conversion.

[0055] Specifically, the constellation point set can be understood as a set of symbols. In digital communication, information is sent in the form of symbols, and each symbol corresponds to a point on the constellation diagram. These symbols are mapped to specific signal waveforms through modulation and transmitted through the channel. That is, the constellation point set can be understood as the set of all possible values corresponding to the target symbols.

[0056] Based on the constellation point set, the bit values at each position of the target symbol can be determined. By accumulating the symbol information with the same bit value at the same position, the bit-level detection information at each bit position can be obtained through conversion. This usually means outputting a value representing its confidence or probability for each bit position, or simply outputting a hard decision result (i.e., 0 or 1).

[0057] Exemplarily, an approximate bit logarithmic likelihood information conversion formula can be adopted to convert the symbol-level logarithmic likelihood information (i.e., symbol-level detection information) into bit-level logarithmic likelihood information (i.e., bit-level detection information): where represents the th constellation point symbol of the modulation scheme, represents the set of constellation points where the nd bit of the symbol in the modulation scheme is 0, represents the set of constellation points where the nd bit of the symbol in the modulation scheme is 1, represents using the bit-level logarithmic likelihood information (i.e., bit-level detection information) as the input of the decoding inner loop.

[0058] Based on any of the above embodiments, through outer loop iteration, the information updated respectively in the detection inner loop and the decoding inner loop is interacted, including: Through decoding inner loop iteration, based on the bit-level detection information, the bit-level decoding information is updated, and the bit-level decoding information is converted into symbol-level decoding information; Through outer loop iteration, based on the difference between the bit-level detection information and the bit-level decoding information, the symbol-level decoding information is updated.

[0059] Specifically, the bit-level detection information is used as the input of the decoding inner loop, and through the decoding inner loop, it is updated to the bit-level decoding information , and the bit-level decoding information is converted into symbol-level decoding information.

[0060] The difference between the bit-level detection information and the bit-level decoding information reflects the accuracy and redundancy of the information during the decoding process, providing guidance for optimizing the decoding algorithm or parameters. Under the outer-loop iteration, the symbol-level decoding information is updated based on the above difference, and the updated symbol-level decoding information may be used in the next decoding inner-loop iteration to further improve the decoding accuracy.

[0061] Exemplarily, the symbol-level decoding information can be updated based on the following formula: where, represents the symbol-level decoding information of the target symbol at the -th iteration, represents the bit-level decoding information of the target symbol at the -th iteration, represents the bit-level detection information of the target symbol at the -th iteration, represents the modulation order, represents the -th bit of the target symbol, represents the indicator function, and the function returns a value of 1 when is satisfied, otherwise 0.

[0062] where the set , and the sign(x) function returns the sign of the input x.

[0063] It should be noted that in the related art, when determining the symbol-level decoding information at the current iteration, it is usually determined based on the following formula: where, represents the symbol-level decoding information of the target symbol at the -th iteration, represents the bit-level decoding information of the target symbol at the -th iteration, represents the bit-level detection information of the target symbol at the -th iteration, represents the modulation order, represents the -th bit of the target symbol.

[0064] Obviously, in the related art, a large number of exponential operations and logarithmic operations , while in the embodiment of the present invention, when determining the symbol-level decoding information, an addition operation of summing is adopted, eliminating a large number of exponential operations and logarithmic operations, greatly reducing the computational complexity and saving the computational time.

[0065] Based on any of the above embodiments, the method further includes: Through outer-loop iteration, after the information updated respectively in the detection inner-loop and the decoding inner-loop is interacted, both the detection inner-loop iteration and the decoding inner-loop iteration use the interacted information as the prior information to assist the information iteration in their respective inner-loops.

[0066] Specifically, during the outer-loop iteration process, the information updated respectively in the detection inner-loop iteration and the decoding inner-loop iteration will be interacted. After this interaction, these information are regarded as the prior information, that is, they are the known information that can be utilized in the subsequent iteration process.

[0067] In the detection inner-loop iteration and the decoding inner-loop iteration, this prior information will be used to assist the information update in their respective loops. Specifically speaking, the prior information can be used as the initial estimate or reference point to help the iteration process converge to the true transmitted information more quickly.

[0068] That is to say, in the embodiment of the present invention, through the information interaction mechanism of the outer-loop iteration, the detection inner-loop iteration and the decoding inner-loop iteration can form a system that works collaboratively. In this system, each inner-loop iteration can utilize the prior information from other iterations, thereby improving its own decoding performance. This collaboration makes the entire iterative decoding process more efficient and accurate.

[0069] Based on any of the above embodiments, the present invention further provides a combined iterative method for detection and decoding, and this method includes: Reconstruct the original posterior update formula in the log-likelihood domain, and the corresponding formula is as follows: Among them, , , respectively represent the posterior information, the symbol-level detection information, and the symbol-level decoding information of the target symbol at the th iteration.

[0070] In the detection inner-loop, the calculation formula of the detection mean value is as follows: Among them, represents the detection mean value at the th iteration. The diag(X) function represents extracting the diagonal elements of the matrix X. Represents the value of the row and column of the channel autocorrelation matrix, represents the diagonal correlation coefficient of the channel autocorrelation matrix of the th received antenna, represents the sum of the received symbols of the channel of the th received antenna, represents the number of transmit antennas.

[0071] In the above formula represents the posterior mean of the th transmit antenna in the i th iteration, where the initialization calculation formula is as follows: Among them, represents the initial value of the posterior mean of the th transmit antenna, , , , represents the channel matrix, represents the noise variance, represents the received signal.

[0072] After obtaining , calculate : Among them, represents the symbol-level detection information of the target symbol in the th iteration, , represents the channel matrix, represents the noise variance, represents the target symbol, represents the detection mean of the th iteration.

[0073] After completing the corresponding calculation, the corresponding posterior information is updated through the detection-decoding outer loop. First, the detection inner loop provides bit log-likelihood information for the decoding inner loop. The symbol-level log-likelihood information can be converted into bit-level log-likelihood information through an approximate bit log-likelihood information conversion formula: Among them, represents the A constellation point symbol, represents the set of constellation points where the th bit of the symbol in the modulation scheme is 0, represents the set of constellation points where the th bit of the symbol in the modulation scheme is 1, represents taking the bit-level log-likelihood information (i.e., bit-level detection information) as the input of the inner decoding loop, and updating it to bit-level log-likelihood information through the inner decoding loop. When passing it back to detection through the detection-decoding outer loop, the conversion from bit-level to symbol-level log-likelihood information needs to be completed. The original formula is as follows: where, represents the symbol-level decoding information of the target symbol at the th iteration, represents the bit-level decoding information of the target symbol at the th iteration, represents the bit-level detection information of the target symbol at the th iteration, represents the modulation order, represents the th bit of the target symbol.

[0074] The embodiment of the present invention approximates the original formula to eliminate logarithmic / exponential operations. The formula is as follows: where, represents the symbol-level decoding information of the target symbol at the th iteration, represents the bit-level decoding information of the target symbol at the th iteration, represents the bit-level detection information of the target symbol at the th iteration, represents the th bit of the target symbol, represents the modulation order, is an indicator function, and the function returns a value of 1 when is satisfied, otherwise 0. For the above set , the sign(x) function returns the sign of the input .

[0075] Obtained by iterative update through the inner loop After that, each data is integrated through the detection - decoding outer loop and updated according to the above posterior update formula , and obtain After that, update the posterior mean : Among them, represents the posterior mean of the th iteration, represents the codeword constraint of the target symbol , which can be understood as the symbol set to which the target symbol belongs, represents the target symbol, represents the th constellation point symbol of the modulation scheme.

[0076] Complete one detection - decoding joint iteration through the above process. The above - mentioned formula for detection and decoding inner loop eliminates matrix inversion and exponentiation based on the log - likelihood domain, and a large number of multiplication operations are also replaced by addition operations, greatly reducing the computational complexity.

[0077] In the embodiment of the present invention, the above - proposed algorithm process eliminates the data dependence that the data in the detection - decoding inner loop needs to update the posterior information through the outer loop successively. As in the posterior update formula, calculating formula and formula shows that the data in each loop can be updated and iterated simultaneously to jointly update the posterior information. Compared with the sequential update of the posterior information, the parallel IDD strategy proposed in the embodiment of the present invention greatly improves the iteration efficiency.

[0078] For the analysis of the specific time complexity, assume that the time required for each detection inner loop and decoding inner loop is and , and the number of times of the detection inner loop, decoding inner loop, and outer loop are , and . Figure 2 is a schematic diagram of the traditional iterative model and the parallel iterative model of the present invention. As Figure 2 shown, the time required for the traditional iterative model to complete all iterations is , and the time required for the parallel iterative model proposed by the present invention to complete all iterations is reduced to .

[0079] Figure 3 is a schematic diagram of the process of another detection - decoding joint iteration method provided by the present invention. As Figure 3 shown, first initialize the posterior mean, and the detection inner loop iteration and the decoding inner loop iteration are executed in parallel. Among them, the number of times of the detection inner loop is , and the number of times of the decoding inner loop is , the number of outer loop iterations is . Under one inner loop iteration of detection, symbol-level detection is obtained . Under one inner loop iteration of decoding, symbol-level decoding is obtained . Then, outer loop iteration is performed, that is, based on and the posterior mean is updated. That is, one inner loop iteration of detection, one inner loop iteration of decoding, and one outer loop iteration constitute one joint iteration of detection and decoding.

[0080] Next, the joint iteration device for detection and decoding provided by the present invention will be described. The joint iteration device for detection and decoding described below can be correspondingly referred to the joint iteration method for detection and decoding described above.

[0081] Based on any of the above embodiments, Figure 4 is a schematic structural diagram of the joint iteration device for detection and decoding provided by the present invention. As Figure 4 shown, the device includes: An inner loop detection module 410, configured to update symbol-level detection information based on channel characteristics under inner loop iteration of detection; An inner loop decoding module 420, configured to update bit-level decoding information based on the codeword constraint of the target symbol under inner loop iteration of decoding; An outer loop module 430, configured to interact the information updated respectively in the inner loop of detection and the inner loop of decoding through outer loop iteration. After the number of outer loop iterations reaches the maximum number of iterations, based on the symbol-level updated detection information and the bit-level updated decoding information corresponding to the maximum number of iterations, the detection and decoding results are determined; Among them, the inner loop iteration of detection and the inner loop iteration of decoding are executed in parallel.

[0082] Based on any of the above embodiments, under inner loop iteration of detection, updating symbol-level detection information based on channel characteristics includes: Under the current inner loop iteration of detection, determine the posterior mean of the transmit antenna corresponding to the target symbol. The posterior mean refers to the estimated expected value of the target symbol; the initial value of the posterior mean is determined based on the noise variance of the received signal and channel characteristics; Update the detection mean based on the posterior mean and channel characteristics; Update the symbol-level detection information based on the updated detection mean and channel characteristics; Convert the updated symbol-level detection information into bit-level detection information.

[0083] Based on any of the above embodiments, the posterior mean is updated based on the following steps: Update the posterior information of the target symbol under the current detection inner loop iteration based on the symbol-level detection information and the symbol-level decoding information, where the posterior information represents the probability or likelihood ratio that the target symbol is correctly decoded, and the initial value of the posterior information is extracted from the received signal; Update the posterior mean based on the updated posterior information and the codeword constraint.

[0084] Based on any of the above embodiments, convert the updated symbol-level detection information into bit-level detection information, including: Based on the bit values at each position of each symbol in the constellation point set, accumulate the symbol information with the same bit value at the same position, and convert it to the bit-level detection information corresponding to the position.

[0085] Based on any of the above embodiments, through the outer loop iteration, interact the information updated in the detection inner loop and the decoding inner loop respectively, including: Through the decoding inner loop iteration, update the bit-level decoding information based on the bit-level detection information, and convert the bit-level decoding information into symbol-level decoding information; Through the outer loop iteration, update the symbol-level decoding information based on the difference between the bit-level detection information and the bit-level decoding information.

[0086] Based on any of the above embodiments, it further includes: After interacting the information updated in the detection inner loop and the decoding inner loop respectively through the outer loop iteration, both the detection inner loop iteration and the decoding inner loop iteration use the interaction information as the prior information to assist the information iteration in their respective inner loops.

[0087] Figure 5 It is a schematic structural diagram of the electronic device provided by the present invention, as Figure 5 shown. The electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540. Among them, the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the logical instructions in the memory 530 to execute the joint detection and decoding iterative method, which includes: under the detection inner loop iteration, update the symbol-level detection information based on the channel characteristics; under the decoding inner loop iteration, update the bit-level decoding information based on the codeword constraint of the target symbol; through the outer loop iteration, interact the information updated in the detection inner loop and the decoding inner loop respectively, until after the number of outer loop iterations reaches the maximum number of iterations, based on the symbol-level updated detection information and the bit-level updated decoding information corresponding to the maximum number of iterations, determine the detection and decoding result; where the detection inner loop iteration and the decoding inner loop iteration are executed in parallel.

[0088] In addition, when the logical instructions in the above-mentioned memory 530 can be implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, 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. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0089] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the detection and decoding joint iterative method provided by the above-mentioned various methods. The method includes: under the detection inner-loop iteration, updating symbol-level detection information based on channel characteristics; under the decoding inner-loop iteration, updating bit-level decoding information based on the codeword constraint of the target symbol; through outer-loop iteration, interacting the information updated respectively in the detection inner-loop and the decoding inner-loop until the number of outer-loop iterations reaches the maximum number of iterations, and then updating the symbol-level detection information and the bit-level decoding information corresponding to the maximum number of iterations to determine the detection and decoding result; wherein, the detection inner-loop iteration and the decoding inner-loop iteration are executed in parallel.

[0090] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the detection and decoding joint iterative method provided by the above-mentioned various methods. The method includes: under the detection inner-loop iteration, updating symbol-level detection information based on channel characteristics; under the decoding inner-loop iteration, updating bit-level decoding information based on the codeword constraint of the target symbol; through outer-loop iteration, interacting the information updated respectively in the detection inner-loop and the decoding inner-loop until the number of outer-loop iterations reaches the maximum number of iterations, and then updating the symbol-level detection information and the bit-level decoding information corresponding to the maximum number of iterations to determine the detection and decoding result; wherein, the detection inner-loop iteration and the decoding inner-loop iteration are executed in parallel.

[0091] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0092] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The 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 for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements 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 detection and decoding joint iterative method, characterized in that: include: Under the detection inner loop iteration, the symbol-level detection information is updated based on the channel characteristics; Under the decoding inner loop iteration, the bit-level decoding information is updated based on the codeword constraint of the target symbol; Through the outer loop iteration, the information updated in the detection inner loop and the decoding inner loop are interacted respectively until the number of iterations of the outer loop reaches the maximum number of iterations, and the detection information is updated at the symbol level and the decoding information is updated at the bit level corresponding to the maximum number of iterations to determine the detection and decoding result; The detection inner loop iteration and the decoding inner loop iteration are performed in parallel.

2. The detection and decoding joint iterative method according to claim 1, characterized in that: The updating of the symbol-level detection information based on the channel characteristics in the detection inner loop iteration includes: Under the current detection inner loop iteration, determine the a posteriori mean of the transmitting antenna corresponding to the target symbol, the a posteriori mean refers to the estimated expected value of the target symbol; the initial value of the a posteriori mean is determined based on the noise variance of the received signal and the channel characteristics; Based on the posterior mean and the channel characteristic, updating the detection mean; Based on the updated detection mean and the channel characteristics, updating the symbol-level detection information; The updated symbol-level detection information is converted into bit-level detection information.

3. The detection and decoding joint iterative method according to claim 2, characterized in that: The posterior mean is updated based on the following steps: Based on the symbol-level detection information and the symbol-level decoding information, updating the a posteriori information of the target symbol under the current detection inner loop iteration, the a posteriori information indicating the probability or likelihood ratio that the target symbol is correctly decoded, and the initial value of the a posteriori information is extracted from the received signal; The posterior mean is updated based on the updated posterior information and the codeword constraint.

4. The detection and decoding joint iterative method according to claim 2, characterized in that: The step of converting the updated symbol-level detection information into bit-level detection information comprises: Based on the bit value of each symbol at each position in the constellation point set, the symbol information of the same position and the same bit value is accumulated and converted to obtain the bit-level detection information of the corresponding position.

5. The detection and decoding joint iterative method according to claim 2, characterized in that: The method of interacting with the information updated in the detection inner loop and the decoding inner loop respectively through the outer loop iteration includes: By iterating the decoding inner loop, based on the bit-level detection information, the bit-level decoding information is updated, and the bit-level decoding information is converted into symbol-level decoding information; Through the outer loop iteration, the symbol-level decoding information is updated based on the difference between the bit-level detection information and the bit-level decoding information.

6. The detection and decoding joint iterative method according to any one of claims 1 to 5, characterized in that: The method further comprises: After the information updated in the detection inner loop and the decoding inner loop are interacted with each other through the outer loop iteration, the detection inner loop iteration and the decoding inner loop iteration both use the interactive information as prior information to assist the information iteration in their respective inner loops.

7. A detection and decoding joint iteration device, characterized in that: include: A detection inner loop module, used for updating symbol-level detection information based on channel characteristics under the detection inner loop iteration; A decoding inner loop module, used for updating bit-level decoding information based on the codeword constraint of the target symbol under the decoding inner loop iteration; The outer loop module is used to interact with the information updated in the detection inner loop and the decoding inner loop respectively through the outer loop iteration until the number of outer loop iterations reaches the maximum number of iterations, and based on the maximum number of iterations corresponding to the symbol-level update detection information and the bit-level update decoding information, determine the detection decoding result; The detection inner loop iteration and the decoding inner loop iteration are performed in parallel.

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, the detection and decoding joint iterative method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the detection and decoding joint iterative method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the detection and decoding joint iterative method according to any one of claims 1 to 6 is implemented.

Images