A hybrid window decoding method and system for a joint source-channel coding system at any time

Abstract

The application belongs to the technical field of decoding, and discloses a hybrid window decoding method of a joint source channel coding system at any time, which aims to reduce the decoding complexity of the system while maintaining the error correction performance equivalent to the existing decoding scheme. The main method of the JHWD scheme is to implement an adaptive local extended window decoding strategy in the sliding window area, which can effectively reduce the complexity caused by the global extended window decoding and reduce the error propagation caused by the sliding window decoding. The method provides a good decoding scheme for the JSCAC system in a low complexity transmission scene. The method provides a low complexity decoding scheme for the joint source channel coding system at any time, and provides the nature for the popularization of the system and the application in the low complexity communication scene.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of decoding, and particularly relates to a hybrid window decoding method and system for a joint source channel coding system at any moment. BACKGROUND

[0002] Joint source channel coding (JSCC) is a coding and decoding technology that designs source and channel coding and decoding in a mutually dependent manner, and realizes optimal coding and decoding of a communication system as a whole. The JSCC technology can be used in any communication system that violates the Shannon separation coding theorem and needs to consider the source end and the channel end at the same time, such as a communication system limited by power, bandwidth, complexity or time delay, a multi-user shared channel system, and different types of communication environments composed of heterogeneous sources or channels.

[0003] There are several typical coding and decoding schemes for joint source channel coding technology. At the sending end, the channel encoder can provide channel state information to the source encoder for bit allocation and optimization of source coding. On the other hand, the source encoder can also provide auxiliary source side information for unequal error protection (UEP) in the channel coding or modulation process. At the receiving end, the priori information and the posteriori information of the source can be used for iterative decoding between the source decoder and the channel decoder, which is one of the most popular JSCC solutions.

[0004] Fresia first proposed a JSCC system using two LDPC codes (referred to as D-LDPC) to complete source compression and channel error correction in the document "Joint source channel coding" in 2002, and implemented joint BP decoding in a global factor graph model framework including source and channel coding structures. After that, many code types have been introduced into this JSCC system, such as block dual protograph LDPC code (DP-LDPC), dual polarity code (D-Polar), space-coupled LDPC (SC-LDPC) code, and space-coupled repetition accumulation code (SC-RA) and the like.

[0005] In the above solution, the application mainly focuses on a joint source-channel arbitrary time coding system (JSCAC) based on SC-RA code. It is a JSCC coding and decoding scheme combining causal space coupling coding and extended window decoding, which can greatly improve the error correction performance of the traditional JSCC system and is more suitable for high-reliability and low-delay communication scenarios. However, the extended window decoding strategy of the JSCAC system has very high decoding complexity and memory space requirements, which affects the practical application of the JSCAC system. In view of this, the application proposes a joint hybrid window decoding (JHWD) method, which aims to reduce the decoding complexity while maintaining the error correction performance comparable to the traditional extended window decoding.

[0006] Through the above analysis, the problems and defects of the prior art are:

[0007] The extended window decoding strategy of the JSCAC system has very high decoding complexity and memory space requirements, which affects the practical application of the JSCAC system. SUMMARY

[0008] In view of the problems existing in the prior art, the application provides a hybrid window decoding method for a joint source-channel arbitrary time coding system.

[0009] The application is implemented as follows: a hybrid window decoding method for a joint source-channel arbitrary time coding system comprises:

[0010] S1, initial extended window decoding stage:

[0011] The decoding delay t satisfies t = j, 1 <= j <= 3 gamma - 2, and the source sub-matrix channel sub-matrix and the prior log-likelihood information of the source perform joint BP decoding on the first to j-th information blocks, where H sc and H cc represent the original source matrix and channel matrix, the values separated by commas in the subscript brackets represent the selected row and column ranges, respectively, and p represents the source probability; the maximum number of decoding iterations is imax1; after decoding, the source posterior log-likelihood information of the first to j-th information blocks at t-1 time is updated using the decoding result at t time as:

[0012] S2, sliding window decoding stage:

[0013] The decoding delay t satisfies t = j, 3 gamma - 2 < j <= T, where T is the maximum decoding delay;

[0014] S3, according to the source posterior log-likelihood information Hard decision is made and the decoding result is output.

[0015] Further, the sliding window decoding stage includes:

[0016] S2-1, sliding mode of the window:

[0017] When j=3γ-1, the first information block is abandoned to decode, the sliding window moves down γ rows and moves right 1 column in H sc and H cc When j>3γ-1, the j-3γ-2th information block is abandoned to decode, the sliding window moves down 1 row and moves right 1 column in H sc and H cc

[0018] S2-2, the regular decoding stage includes:

[0019] using the source sub-matrix the channel sub-matrix and the source initial likelihood information Z sc The joint BP decoding is performed on the j-3γ+3th to jth information blocks, and the maximum iteration number is imax1; after the decoding, the posterior log-likelihood information of the j-3γ+3th to jth information blocks is updated as:

[0020] S2-3, the modified decoding stage includes:

[0021] The initial value of the accompanying sub-check flag is S_check=1; when 1<=k<=γ, if S_check=1, the source sub-matrix the channel sub-matrix and the updated source posterior log-likelihood information The joint BP decoding is performed on the j-3γ+3th to j-γ+kth information blocks; the maximum iteration number is imax2, and imax2 is smaller than imax1; after the decoding, the accompanying sub-check is performed, if S_check=0, the local adaptive extended window decoding in the sliding window is ended; after the decoding, the posterior log-likelihood information of the j-3γ+3th to j-2γ+2th information blocks is updated as:

[0022] Another object of the present application is to provide a hybrid window decoding system of a joint source channel coding system at any time, which includes:

[0023] The extended window decoding module is used for the initial extended window decoding stage.

[0024] The sliding window decoding module is used for the sliding window decoding stage.​

[0025] The decoding result output module is configured to perform hard decision according to the source posterior log-likelihood information and output the decoding result.

[0026] Another object of the present application is to provide a computer device comprising a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to enable the processor to perform the steps of the hybrid window decoding method of the joint source-channel coding system at any time.

[0027] Another object of the present application is to provide a computer readable storage medium storing a computer program, the computer program being executed by a processor to enable the processor to perform the steps of the hybrid window decoding method of the joint source-channel coding system at any time.

[0028] Another object of the present application is to provide an information data processing terminal for implementing the hybrid window decoding system of the joint source-channel coding system at any time.

[0029] In combination with the above technical solutions and the technical problems solved, the technical solution to be protected by the present application has the following advantages and positive effects:

[0030] The joint source-channel coding system at any time is a kind of coding and decoding system for industrial control, tracking system and other systems with high requirements for low latency and high reliability. However, the complexity of the traditional joint source-channel extended window decoding algorithm increases linearly with the increase of the length of the information sequence, thereby limiting the application and promotion of the joint source-channel coding system at any time. Considering that the information sequence in the low latency and high reliability scenario has the characteristics of limited coupling length, three strategies are mainly adopted to limit the linear growth of decoding complexity: 1) the starting point of the decoding window is movable; 2) the length of the decoding window does not change; 3) the extended window decoding in the sliding window is performed in an adaptive manner. The hybrid window decoding algorithm proposed in the present application provides a low complexity decoding scheme for the joint source-channel coding system at any time while ensuring decoding performance, solves the problems and bottlenecks of the system in decoding complexity, and provides greater possibility for the promotion of the system and the application in the low complexity communication scenario.

[0031] Secondly, the technical solution of the present application fills the technical gap in the industry at home and abroad: In order to ensure the transmission characteristics at any time, i.e., the decoding error rate decreases exponentially with the time delay, the decoding algorithm must use an extended window form in the joint source-channel coding system at any time. The simple sliding window decoding will destroy the transmission characteristics at any time. The technical solution of the present application makes the starting point of the traditional extended decoding window movable, and the window size is no longer unlimitedly increased. Thus, the transmission characteristics at any time unique to the extended window decoding are retained, and the linearly increasing decoding complexity caused by the extended window decoding is limited. Thus, the technical gap in the industry at home and abroad is filled.

[0032] (2) Whether the technical solution of the present application solves the technical problem that people have been eager to solve but have always failed to succeed: Linearly increasing decoding complexity has always been a bottleneck problem faced by the joint source-channel coding system at any time, and is also a key factor that makes it difficult for the system to be applied in practice. The technical solution of the present application effectively reduces the decoding complexity of the system while ensuring decoding performance, making the industrial application of the joint source-channel coding system at any time no longer out of reach.

[0033] (3) Whether the technical solution of the present application overcomes technical bias: The industry has always believed that sliding window decoding cannot guarantee the system to have transmission characteristics at any time. The technical solution of the present application adaptively implements extended window decoding within the sliding window, so that the decoding performance still maintains the transmission characteristics at any time, thereby overcoming the technical bias.

[0034] Thirdly, the technical solution of the present application solves the problems of low coding efficiency, high error rate and insufficient energy utilization in the source-channel coding system in the prior art in industrial applications. By introducing the hybrid window decoding method, the present application can jointly process the source and channel coding at any time, greatly improving the decoding accuracy and data transmission efficiency.

[0035] Specifically, the significant technical progress of the present application includes:

[0036] 1. Improve coding efficiency: By jointly processing the source sub-matrix and channel sub-matrix in the initial extended window and sliding window decoding, the utilization of energy and information in the decoding process is optimized, significantly improving the coding efficiency in the data transmission process, especially suitable for scenarios that require low delay and high efficiency transmission.

[0037] 2. Reduce error rate: The decoding strategy based on the posterior log-likelihood information is adopted, so that the system can adjust according to the real-time state of the received signal in a dynamic transmission environment, effectively reducing the error rate in the decoding process and enhancing the reliability of the system.

[0038] 3. Adaptive decoding window adjustment: The sliding window mechanism of the present invention combines local adaptive extension decoding, enabling the system to flexibly adjust the size and position of the decoding window when facing different channel conditions, effectively dealing with burst interference of the channel and further improving the robustness of the system.

[0039] 4. Enhanced application flexibility: The system can be widely applied in communication networks, satellite transmission, Internet of Things, etc., and through its efficient encoding and decoding mechanism, it can significantly improve the quality and speed of information transmission in various complex scenarios.

[0040] The present invention solves the bottleneck problem in the existing source channel coding system in practical industrial applications, achieves significant technical progress, and helps to improve the overall performance of modern communication systems. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a hybrid window decoding method flowchart of the joint source channel coding system at any moment provided by the embodiment of the present invention.

[0042] Figure 2 is a hybrid window decoding system structure block diagram of the joint source channel coding system at any moment provided by the embodiment of the present invention.

[0043] Figure 3 is a hybrid window decoding structure diagram of the joint source channel coding (JSCAC) system provided by the embodiment of the present invention.

[0044] Figure 4 is a density evolution-based asymptotic analysis diagram of the decoding complexity of the JSCAC system provided by the embodiment of the present invention.

[0045] Figure 5 is a density evolution-based asymptotic analysis diagram of the characteristics at any moment of the JSCAC system provided by the embodiment of the present invention.

[0046] Figure 6 is a graph of the error correction performance of any information block in the data chain with decoding delay change provided by the embodiment of the present invention.

[0047] Figure 7 is a graph of the overall bit error rate performance of the data chain and decoding complexity when the decoding delay is t = 100 provided by the embodiment of the present invention.

[0048] Figure 8 is a graph of the influence of the length of the information block on the bit error rate performance of the data chain and decoding complexity provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0050] The joint source-channel arbitrary-time coding system provided in this embodiment of the invention includes:

[0051] The source coding module is used to perform source coding on the input source information and output the source-coded information.

[0052] The channel coding module is used to perform channel coding on the source-encoded information to generate encoded channel information;

[0053] The hybrid window decoding module includes an initial extended window decoding stage and a sliding window decoding stage. The initial extended window decoding is used to perform joint BP decoding on the information block in the time delay range of 1 to 3γ-2, and the sliding window decoding is used to perform joint BP decoding by moving the window in the time delay range of 3γ-2 to T.

[0054] The hard decision module is used to make hard decisions based on the posterior log-likelihood information of the information source and output the decoding results.

[0055] The hybrid window decoding module further includes:

[0056] The source posterior log-likelihood information update module is used to update the source posterior log-likelihood information of the information block at the previous time step based on the result at the current decoding time step after the initial extended window decoding stage is completed.

[0057] The sliding window decoding module includes:

[0058] The window sliding control unit is used to control the sliding mode of the window in the source matrix and the channel matrix. When the time delay reaches 3γ-1, the decoding of the first information block is abandoned, and the window is slid down γ rows and shifted to the right by one column. When the time delay exceeds 3γ-1, j-3γ-2 information blocks are abandoned, and the window is shifted down one row and shifted to the right by one column.

[0059] The sliding window decoding module further includes:

[0060] The accompanying sub-check module is used to perform accompanying sub-checks on the information blocks from j-3γ+3 to j-γ+k during the correction decoding stage of sliding window decoding, and decide whether to continue the local adaptive expansion window decoding process based on the check results.

[0061] like Figure 1 As shown in the figure, the hybrid window decoding method for a joint source-channel arbitrary-time coding system provided by the present invention includes the following steps:

[0062] S1, initial extended window decoding stage:

[0063] The decoding delay t satisfies t = j, 1 <= j <= 3 gamma - 2, and the source sub-matrix The channel sub-matrix and the prior log-likelihood information of the source Joint BP decoding is performed on the first to j-th information blocks, where gamma is the coupling length, H sc and H cc represent the original source matrix and the channel matrix, the numbers separated by commas in the subscript brackets represent the selected row and column ranges, respectively, and p represents the source probability; the maximum number of iterations for decoding is imax1; after decoding, the source posterior log-likelihood information of the first to j-th information blocks at t-1 time is updated using the decoding result at t time as follows:

[0064] S2, sliding window decoding stage:

[0065] The decoding delay t satisfies t = j, 3 gamma - 2 < j <= T, where T is the maximum decoding delay;

[0066] S3, hard decision is made according to the source posterior log-likelihood information to output the decoding result.

[0067] The sliding window decoding stage provided by the embodiment of the application comprises:

[0068] S2-1, sliding mode of the window:

[0069] When j = 3 gamma - 1, the decoding of the first information block is abandoned, and the sliding window is moved down by gamma rows and right by one column in H sc and H cc ; when j > 3 gamma - 1, the decoding of the j-3 gamma - 2-th information block is abandoned, and the sliding window is moved down by one row and right by one column in H sc and H cc ;

[0070] S2-2, conventional sliding window decoding stage:

[0071] The source sub-matrix The channel sub-matrix and the initial likelihood information Z of the source sc Joint BP (Belief Propagation) decoding is performed on the j-3 gamma + 3-th to j-th information blocks, and the maximum number of iterations is imax1; after decoding, the posterior log-likelihood information of the j-3 gamma + 3-th to j-th information blocks is updated as follows:

[0072] S2-3, adaptive local extended window decoding stage:

[0073] Let the initial value of the accompanying sub-check flag be S_check=1;When 1<=k<=γ, if S_check=1, use the source sub-matrix Channel sub-matrix And the updated source posterior log-likelihood information Perform joint BP decoding on the j-3γ+3~j-γ+k information blocks;The maximum iteration number is imax2, and imax2 is smaller than imax1;After decoding, the accompanying sub-check is performed, and if the result of the accompanying sub-check is S_check=0, the local adaptive expansion window decoding in the sliding window is ended;After decoding, the posterior log-likelihood information of the j-3γ+3~j-2γ+2 information blocks is updated as follows: The calculation formula of the accompanying sub-check is as follows:

[0074]

[0075] Wherein And The source decoding result and the channel decoding result of the kth joint BP decoding result are respectively. It is the source parity check matrix involved in the kth joint BP decoding.

[0076] S3, according to the source posterior log-likelihood information Hard decision is performed, and the decoding result is output.

[0077] As Figure 2 The embodiment of the present application provides a mixed window decoding system of a joint source channel coding system at any moment, which comprises:

[0078] An expansion window decoding module is used for the initial expansion window decoding stage.

[0079] A sliding window decoding module is used for the sliding window decoding stage.

[0080] A decoding result output module is used for performing hard decision according to the source posterior log-likelihood information, and outputting the decoding result.

[0081] Another purpose of the present application is to provide a computer device, which comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor, so that the processor executes the steps of the mixed window decoding method of the joint source channel coding system at any moment.

[0082] Another object of the present application is to provide a computer readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the hybrid window decoding method for the joint source-channel arbitrary-time encoding system.

[0083] Another object of the present application is to provide an information data processing terminal for implementing the hybrid window decoding system for the joint source-channel arbitrary-time encoding system.

[0084] The present application is embodied as follows:

[0085] As shown in the decoding method, a low-complexity decoding scheme is provided for the joint source-channel arbitrary-time encoding system based on SC-RA code. Figure 3

[0086] The technical solution of the present application is as follows:

[0087] A decoding method for a SC-RA code-based encoding and decoding system, which is a joint source-channel arbitrary-time encoding system. The method is suitable for a scenario where the chain length L and the coupling length γ satisfy L→∞ and L>γ. The size of the decoding window is selected to include the coupling length interval where the information block to be decoded is located, and the two adjacent forward coupling length intervals, a total length of 3γ-2 information blocks, which are referred to as the first coupling interval, the second coupling interval and the third coupling interval, respectively. The decoding method includes the following steps:

[0088] S1, initial extended window decoding stage: the decoding delay t satisfies t=j, 1<=j<=3γ-2

[0089] Use the source sub-matrix Channel sub-matrix And the prior log-likelihood information of the source Perform joint BP decoding on the first to j-th information blocks, where H sc And H cc Indicate the original source matrix and channel matrix, the values separated by commas in the brackets of the subscripts represent the selected row and column ranges, respectively, and p represents the source probability; the maximum number of iterations for decoding is imax1. After decoding, the source posterior log-likelihood information of the first to j-th information blocks at time t-1 is updated using the decoding result at time t as follows:

[0090] S2, sliding window decoding stage: the decoding delay t satisfies t=j, 3γ-2<j<=T, where T is the maximum decoding delay

[0091] S2-1, sliding mode of the window: when j=3γ-1, the first information block is abandoned for decoding, and the sliding window is in H sc And H​cc corresponding to the (j-3y-2)th information block is discarded, and the sliding window moves down 1 row and right 1 column in H sc and H cc corresponding to the (j-3y-2)th information block is discarded, and the sliding window moves down 1 row and right 1 column in H

[0092] S2-2, the regular sliding window decoding stage: using the source sub-matrix channel sub-matrix

[0093]

[0094] and the updated source a posteriori log-likelihood information Zsc the joint BP decoding is performed on the (j-3y+3)th to jth information blocks, and the maximum iteration number is imaxl. After the decoding, the a posteriori log-likelihood information of the (j-3y+3)th to jth information blocks is updated as:

[0095] S2-3, the adaptive local extended window decoding stage: the initial value of the accompanying sub-check flag is set as S_check = 1; when 1 <= k <= y, if S_check = 1, the source sub-matrix channel sub-matrix and the updated source a posteriori log-likelihood information the joint BP decoding is performed on the (j-3y+3)th to (j-2y+2)th information blocks; the maximum iteration number is imax2, and imax2 is smaller than imaxl. After the decoding, the accompanying sub-check is performed, and if S_check = 0, the local adaptive extended window decoding in the sliding window is ended. After the decoding, the a posteriori log-likelihood information of the (j-3y+3)th to (j-2y+2)th information blocks is updated as:

[0096] S3, the hard decision is performed according to the source a posteriori log-likelihood information to output the decoding result

[0097] The hybrid window decoding algorithm proposed in the present application is marked as (joint hybrid widow decoding, JHWD) in the embodiment description and the accompanying drawings, and the comparative cases are the joint expanding window decoding algorithm (joint expanding window decoding, JEWD) and the partial joint expanding window decoding algorithm (partial joint expanding window decoding, PJEWD) in the document "Joint Source Channel Anytime Coding". The numerical analysis method used in the embodiment is the density evolution asymptotic analysis method proposed in the document "Joint Source Channel Anytime Coding based on Spatially Coupled Repeat-Accumulate Codes" for analyzing the performance of the SC-RA based JSCAC system.

[0098] I. The specific application field or related product of the present application.

[0099] The specific application field of the present application is the decoder or output player in a low-latency high-reliability wireless communication system, a control system, and a tracking system, and there is no related product on the market.

[0100] II. The related evidence of the technical effects obtained by the embodiments of the present application.

[0101] The parameter settings of the embodiments are as follows: the degrees of the source and channel variable nodes are Q sc =4 and Q cc =3, respectively; the degrees of the source and channel check nodes are A sc =12 and A cc =3, respectively; the exponential connection probabilities of the source and channel are λ sc =0.5 and λ cc =0.1, respectively; the coupling lengths of the source and channel are both set to γ=5; the maximum iteration numbers of decoding at each stage are i max1 =30, i max2 =5, respectively; the source probability is p=0.02; and the signal-to-noise ratio of the channel is 2dB.

[0102] Embodiment:

[0103] In this example, numerical analysis and simulation are used to evaluate the decoding complexity and decoding performance of the JHWD algorithm proposed in the present application, wherein Figure 4 and Figure 5 are the numerical analysis results, Figure 6 to Figure 8 is the simulation result.

[0104] Figure 4 The figure shows the average information update times of three decoding schemes with respect to decoding delay t. When decoding delay t is small, the average information update times of three schemes are similar. When t increases, the average information update times of three schemes increase linearly. Figure 4 The right Y axis of the figure shows the reduction ratios of average information update times of PJEWD decoding scheme and JHWD decoding scheme of the present application with respect to the conventional JEWD scheme. The reduction ratios of two schemes increase rapidly in the small interval of t<600, and then the growth rate tends to be flat when t>600. In the decoding delay interval of t=100-1000, the reduction ratio of average information update times of PJEWD with respect to JEWD increases from 26% to 48%, while the reduction ratio of average information update times of JHWD decoding scheme of the present application with respect to JEWD increases from 58% to 94%. That is, the reduction ratio of average information update times of JHWD decoding scheme of the present application is about 2 times higher than that of PJEWD scheme.

[0105] Figure 5 The figure shows the decoding error probability of any information block in the data chain with respect to decoding delay t. Since any information block in the data chain has the same decoding characteristics at any time, the characteristics curves of adjacent information blocks of each scheme are shown for easy observation and analysis. From the figure, it can be seen that the decoding error probability curves of JEWD and PJEWD schemes have similar decreasing speed at the same coupling length (γ=5); while the decreasing speed of error probability curve of JHWD scheme of the present application is relatively slow, because the decreasing speed of error probability of JHWD scheme is related to the size of window, and when γ=5, the window length is 3γ-2=13. Figure 5 At the same time, the error probability curves of JHWD scheme at different coupling lengths are shown. It can be seen that the larger the coupling length, the longer the window size, the smaller the slope of error probability curve, i.e. the slower the decreasing speed, but a lower decoding error probability can be obtained. Correspondingly, it can be seen from the figure that the larger the coupling length, the higher the decoding complexity. Figure 5 Figure 5

[0106] Figure 6 The figure shows the decoding error probability of any information block in the data chain with respect to decoding delay t. Since any information block in the data chain has the same decoding characteristics at any time, the characteristics curves of adjacent information blocks of each scheme are shown for easy observation and analysis. From the figure, it can be seen that the decoding error probability curves of JEWD and PJEWD schemes have similar decreasing speed at the same coupling length (γ=5); while the decreasing speed of error probability curve of JHWD scheme of the present application is relatively slow, because the decreasing speed of error probability of JHWD scheme is related to the size of window, and when γ=5, the window length is 3γ-2=13. Figure 6 ​​It can be seen that the bit error rate (BER) performance of the D-LDPC based JSCC scheme in the document "Joint source channel coding" cannot be improved with the increase of decoding delay, and thus is a straight line parallel to the horizontal axis. The BER performance of the JSCAC system with causal coupling structure has the opportunity to be further improved at subsequent decoding delay time. When γ = 5, it is found that the BER curve of the JHWD scheme proposed in the present application is located between the BER curves of the JEWD and PJEWD schemes, the BER curve of the PJEWD scheme has the largest slope, and the BER curves of the PJEWD and JEWD schemes reach a flat level after t = 10, while the JHWD scheme proposed in the present application reaches a flat level at the latest (t = 13, i.e. the window length), which is consistent with the numerical analysis result of Figure 6 . Moreover, the error floor of the JHWD scheme proposed in the present application is closer to that of the PJEWD scheme. When the coupling length is increased to γ = 8, the BER curve of the JHWD scheme proposed in the present application has a smaller slope than that when γ = 5, but can obtain a lower error probability. At t = 8, the BER value decreases by less than 10-6. The experiment shows that the decoding characteristics of the JHWD scheme proposed in the present application at any time are between those of the JEWD and PJEWD schemes.

[0107] Figure 7 The overall BER performance of the data link and decoding complexity analysis when the information block length is Ncc = 32 bits and t = 100 are observed. In terms of BER performance, Figure 7 (a), the JHWD scheme proposed in the present application has a smaller curve slope than the PJEWD and JEWD schemes, and performs worst in the low SNR stage, and is closer to the BER performance of the JEWD scheme. In the high SNR stage greater than 1.5 dB, the JHWD scheme can obtain a lower error floor than the JEWD scheme, and is closer to the BER performance of the PJEWD scheme. In terms of decoding complexity, Figure 7(b) shows the normalized average information update times of JHWD and PJEWD schemes relative to JEWD scheme. The decoding complexity of traditional JEWD scheme does not change with the change of SNR, so it is a parallel line. The decoding complexity of JHWD and PJEWD schemes is reduced relative to JEWD scheme, and the performance is more obvious in the high SNR stage. When SNR = 3 dB, the average information update times of PJEWD scheme is reduced by about 26.47% compared with JEWD scheme, and the average information update times of JHWD scheme proposed by the application is reduced by about 69.61% compared with JEWD scheme. The experiment shows that the JHWD scheme proposed by the application can greatly reduce the decoding complexity while maintaining similar BER performance to the traditional JEWD scheme and PJEWD scheme.

[0108] Figure 8 The influence of different information block lengths Ncc on the overall decoding performance and decoding complexity of the data chain under the same data chain length is observed. It is known from Figure 8 (b) that the larger the information block length, the lower the decoding complexity of the JHWD scheme; and as shown in Figure 8 (a), in terms of BER performance, increasing the block length does not bring about a sustained decline in BER performance. When the information block length Ncc is greater than 128 bits, the larger block length causes the data chain to have fewer opportunities for repeated error correction in the low SNR region, resulting in a decline in BER performance. When the SNR is greater than 1.5 dB, the better channel environment reduces the performance difference between it and the shorter block length scheme. On the other hand, under the same coupling length and window length conditions, the increase in the information block length will result in an increase in the proportion of initial extended window decoding and a decrease in the proportion of sliding window decoding in the JHWD scheme, thereby improving the BER performance of the JHWD scheme in the high SNR region and even exceeding the shorter block length scheme. Therefore, under the premise of known observation data chain length, careful selection of the length of the information block can obtain performance gain in both error correction performance and decoding complexity.

[0109] It should be noted that embodiments of the present application can be realized by hardware, software, or a combination of software and hardware. The hardware portion can be realized by a special logic; the software portion can be stored in a memory and executed by a proper instruction execution system, such as a microprocessor or a specially designed hardware. A person of ordinary skill in the art can understand that the above-mentioned apparatus and method can be realized by computer executable instructions and / or included in processor control codes, such as a carrier medium, such as a magnetic disk, CD or DVD-ROM, a programmable memory, such as a read-only memory (firmware), or a data carrier, such as an optical or electronic signal carrier. The apparatus of the present application and its modules can be realized by a hardware circuit, such as a very large scale integrated circuit or a gate array, a semiconductor, such as a logic chip, a transistor, or a programmable hardware device, such as a field programmable gate array, a programmable logic device, or the like, by software executed by various types of processors, or by a combination of the above-mentioned hardware circuit and software, such as firmware.

[0110] The above description is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any modification, equivalent replacement, and improvement within the technical range disclosed by the present application, and within the spirit and principle of the present application, should be included in the protection scope of the present application.

Claims

1. A hybrid window decoding method for a joint source-channel coding system at any time instant, characterized in that, The method comprises the following steps: S1, a mixed window decoding stage, comprising an initial extended window decoding stage and a sliding window decoding stage, wherein the initial extended window decoding is used for jointly BP decoding the information blocks when the delay range is 1 to 3γ-2, and the sliding window decoding is used for jointly BP decoding through moving the window when the delay range is 3γ-2 to T; S2, log-likelihood information of the source posterior perform hard decision and output the decoding result; The initial extended window decoding stage comprises: The decoding delay t satisfies t = j, 1 <= j <= 3γ-2, and the source sub-matrix is used The channel sub-matrix The prior log-likelihood information of the source Joint BP decoding is performed on the first to j-th information blocks, where H sc And H cc The original source matrix and channel matrix are represented by the numbers separated by commas in the subscript brackets, the selected row and column ranges are represented by the numbers separated by commas in the subscript brackets, and p represents the source probability; the maximum number of iterations for decoding is imax1; after decoding, the source posterior log-likelihood information of the first to j-th information blocks at t-1 time is updated using the decoding result at t time: The sliding window decoding stage comprises: The decoding delay t satisfies t = j, 3γ-2 < j <= T, wherein T is the maximum decoding delay; a window sliding control unit is used to control the sliding mode of the window in the source matrix and the channel matrix, when the delay reaches 3γ-1, the decoding of the first information block is abandoned, and the window is slid downward by γ rows and rightward by one column; when the delay exceeds 3γ-1, j-3γ-2 information blocks are abandoned, and the window is slid downward by one row and rightward by one column; An accompanying sub-check module is used to perform accompanying sub-check on the j-3γ+3 to j-γ+k information blocks in the correction decoding stage of the sliding window decoding, and whether to continue the decoding process of the locally adaptive extended window is determined according to the check result.

2. The hybrid window decoding method of claim 1, wherein the step of decoding the combined source and channel encoded signal comprises the steps of: determining a first window size; determining a second window size; and decoding the combined source and channel encoded signal using the first window size and the second window size. The sliding window decoding stage comprises: S2-1, the sliding mode of the window: When j = 3y - 1, the first information block is skipped, the sliding window is shifted down by y rows and right by one column in H sc and H cc When j = 3y - 1, the first information block is skipped, the sliding window is shifted down by y rows and right by one column in H sc and H cc When j = 3y - 1, the first information block is skipped, the sliding window is shifted down by y rows and right by one column in H S2-2, a normal decoding stage: Using source sub-matrices Channel sub-matrices And source initial likelihood information Z sc Perform joint BP decoding on the j-3γ+3~j information blocks, with a maximum number of iterations imax1; after decoding, update the posterior log-likelihood information of the j-3γ+3~j information blocks as: S2-3, a correction decoding stage: Let the initial value of the accompanying sub-check flag be S_check = 1; when 1 <= k <= γ, if S_check = 1, use the source sub-matrix Channel sub-matrix And the updated source posterior log-likelihood information Perform joint BP decoding on the j-3γ+3~j-γ+k information blocks; the maximum iteration number is imax2, and imax2 is smaller than imax1; after decoding, perform accompanying sub-checking, if S_check = 0, end the local adaptive extended window decoding in the sliding window; after decoding, update the posterior log-likelihood information of the j-3γ+3~j-2γ+2 information blocks as follows:

3. A hybrid window decoding system for implementing the decoding method of any one of claims 1-2 in a joint source-channel coding system at any time instant, characterized by The method comprises the following steps: A source encoding module is used to source encode the input source information and output the source encoded information; A channel encoding module is used to channel encode the source encoded information to generate the encoded channel information; A mixed window decoding module comprises an initial extended window decoding module and a sliding window decoding module, wherein the initial extended window decoding module is used for jointly BP decoding the information blocks when the delay range is 1 to 3γ-2, and the sliding window decoding module is used for jointly BP decoding through moving the window when the delay range is 3γ-2 to T; A hard decision module is used to perform hard decision on the source posterior log-likelihood information and output the decoding result.

4. The joint source-channel arbitrary-time encoding system of claim 3, wherein, The mixed window decoding module further comprises: A source posterior log-likelihood information updating module is used to update the source posterior log-likelihood information of the information blocks at the previous time based on the result at the current decoding time after the initial extended window decoding stage ends.

5. The joint source-channel arbitrary-time encoding system of claim 3, wherein, The sliding window decoding module comprises: A window sliding control unit is used to control the sliding mode of the window in the source matrix and the channel matrix, when the delay reaches 3γ-1, the decoding of the first information block is abandoned, and the window is slid downward by γ rows and rightward by one column; when the delay exceeds 3γ-1, j-3γ-2 information blocks are abandoned, and the window is slid downward by one row and rightward by one column.

6. The joint source-channel arbitrary-time encoding system of claim 3, wherein, The sliding window decoding module further comprises: An accompanying sub-check module is used to perform accompanying sub-check on the j-3γ+3 to j-γ+k information blocks in the correction decoding stage of the sliding window decoding, and whether to continue the decoding process of the locally adaptive extended window is determined according to the check result.

7. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the mixed window decoding method of the joint source channel decoding system at any time according to any one of claims 1-2.

8. A computer readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the hybrid window decoding method of a joint source-channel coding system at any time instant as claimed in any one of claims 1-2.

9. An information data processing terminal, characterized by The information data processing terminal is configured to implement the hybrid window decoding system of a joint source-channel coding system at any time instant as claimed in claim 3.

Images