Forward metric normalization method and system for CCSDS131.0-B-3 standard Turbo decoding, electronic equipment and storage medium

Abstract

The invention provides a forward measurement normalization method and system for CCSDS131.0-B-3 standard Turbo decoding, electronic equipment and a storage medium, and the method and system are used for carrying out the systematic simulation analysis of the value change behavior of forward measurement Alpha of each state in the decoding process of a CCSDS131.0-B-3 standard Turbo code frame. Obtaining statistical characteristics of each state which becomes a forward measurement Alpha maximum value state in a forward recursion process; the method comprises the following steps: based on statistical characteristics, screening out a fixed state from all states of a CCSDS131.0-B-3 standard Turbo code grid diagram as a reference state for forward measurement Alpha normalization in the whole decoding process; when recursive calculation of forward measurement Alpha is carried out by a Log-MAP algorithm, for each moment t of a decoding process, a measurement value of a reference state at the current moment t is subtracted from non-normalized forward measurement values of all states at the current moment, so that normalization operation of the forward measurement Alpha of all states at the moment is completed.

Description

Technical Field

[0001] This invention belongs to the field of channel coding and digital communication technology, and particularly relates to a forward metric normalization method, system, electronic device and storage medium for Turbo decoding of CCSDS131.0-B-3 standard. Background Technology

[0002] Turbo codes, due to their superior error correction performance approaching the Shannon limit, have been adopted as one of the core channel coding schemes in the CCSDS (Advisory Committee on Space Data Systems) standard and are widely used in space communication links such as low-Earth orbit satellites and deep space probes. The log-MAP algorithm, a logarithmic domain variant of the BCJR (Bahl, Cocke, Jelinek, and Raviv) algorithm, is the core algorithm for implementing maximum a posteriori probability decoding of Turbo codes. During the forward recursion of this algorithm, the calculated forward metric (usually called the alpha metric) continuously accumulates and increases with the number of recursion steps. In hardware implementations with finite word lengths, this can easily lead to data overflow, causing decoding failure. Therefore, a normalization operation must be introduced during the forward metric iteration process to control its numerical range.

[0003] Existing normalization methods are mainly divided into three categories. The first category is the "dynamic maximum value finding method," which compares the forward metric values ​​of all states in real time at each recursive time step, finds the maximum value, and then subtracts this maximum value from the metric values ​​of all states. Although this method is direct and effective, it requires completing an N-way comparator tree (N is the number of grid states, N=16 in the CCSDS standard) in each clock cycle, resulting in high hardware complexity (O(N)) and long critical path delay, which severely restricts the decoder's operating clock frequency and throughput, becoming a bottleneck for high-speed decoding implementation. The second category is the "threshold-triggered normalization method," which presets a threshold for the metric values. When one or more metric values ​​exceed the threshold, a subtraction operation on all metrics is triggered. However, the setting of the threshold lacks universal theoretical guidance and often relies on simulation testing of specific scenarios or complex dynamic adjustment mechanisms. This not only introduces additional design overhead but may also lead to untimely normalization (still with the risk of overflow) or excessive normalization (increasing unnecessary operations, which may affect numerical accuracy) due to improper threshold settings. The third type is the "modulo normalization method," which utilizes the modulo 2 of the two's complement. n This feature allows overflowing values ​​in the forward metric recursion to wrap naturally, maintaining the correctness of the relative difference. However, this method requires a sufficiently large bit width for the metric to maintain an effective dynamic range, increasing the consumption of hardware storage resources.

[0004] In summary, existing Turbo decoding forward metric normalization methods are insufficient in terms of hardware complexity, timing performance, and design simplicity, making it difficult to simultaneously meet the high reliability and real-time performance requirements of the CCSDS131.0-B-2 standard, as well as the stringent requirements of low power consumption and small size for spacecraft equipment. Therefore, there is an urgent need for a normalization method that can guarantee decoding performance while having extremely low hardware implementation complexity. Summary of the Invention

[0005] This invention provides a forward metric normalization method, system, electronic device, and storage medium for Turbo decoding of the CCSDS131.0-B-3 standard.

[0006] The technical solution of this invention is: a forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard, comprising the following steps: S1: Conduct a systematic simulation analysis of the numerical change behavior of the forward metric Alpha in each state during the decoding process of the CCSDS131.0-B-3 standard Turbo code frame, and obtain the statistical characteristics of each state becoming the state with the maximum value of the forward metric Alpha during the forward recursion process. S2: Based on the statistical characteristics, select a fixed state from all states of the CCSDS131.0-B-3 standard Turbo code trellis diagram as the reference state for forward metric Alpha normalization in the entire decoding process; S3: When the Log-MAP algorithm performs recursive calculation of the forward metric Alpha, for each time t in the decoding process, the unnormalized forward metric values ​​of all states at the current time are subtracted from the metric value of the base state at the current time t, thereby completing the normalization operation of the forward metric Alpha of all states at that time.

[0007] Preferably, the statistical features include a first statistic and a second statistic; The first statistic is configured as the probability P_max(s) of each state of the CCSDS131.0-B-3 standard Turbo code trellis graph becoming the state with the forward metric alpha maximum during the decoding of a complete frame of data, where s represents the state index; The second statistic is configured to be the statistical distribution characteristic of the time interval Δt experienced between two consecutive states that are the maximum values ​​of the forward metric Alpha for each state.

[0008] Preferably, the second statistic includes the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval Δt; In step S2, a fixed state is selected as the reference state for the forward metric Alpha normalization throughout the entire decoding process. The selection criteria for the reference state are as follows: Select a state s0, and limit the probability P_max(s0) of s0 becoming the maximum value state to be higher than a preset first threshold, and the standard deviation σ_Δt(s0) of the time interval Δt between s0 becoming the maximum value state to be lower than a preset second threshold.

[0009] Preferably, in step S2, a fixed state is selected as the reference state for the forward metric Alpha normalization throughout the entire decoding process, including the following steps: S21: For the Turbo code trellis diagram of CCSDS131.0-B-3 standard, Monte Carlo simulation is performed by inputting deterministic test sequences and randomized input sequences under the frame length defined in the standard. Calculate the probability P_max(s) of each state becoming the maximum state of the forward metric Alpha under the two input sequences, and calculate the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval sequence in which each state becomes the maximum state; The state that exhibits both probability frequency and time stability that meet the first threshold requirement under both input sequences is selected as the baseline state. The reference state is defined as state 4 or state 13 out of the 16 state codes for the CCSDS131.0-B-3 standard Turbo code trellis diagram.

[0010] Preferably, after the reference state is selected, it is fixed and applied to the CCSDS131.0-B-3 standard Turbo decoding task.

[0011] Preferably, in S3, the Log-MAP algorithm is used to recursively calculate the forward metric Alpha. The mathematical expression for the recursive calculation is: A t+1 (v) = max(A) t (u) - A t (s0) + Γ t (u, v) Where u is a grid state at time t, v is a valid grid state connected to u at time t+1, and (u,v) represents a valid state transition relationship. A t (u) is the normalized forward metric of the state u at time t, A t (s0) is the normalized forward metric value of the reference state at time t, Γ t (u, v) is the branch metric for the transition from state u to state v at time t.

[0012] Based on the same concept, this invention also provides a forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding, for performing the forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in any of the above-mentioned methods, including: The statistical characteristic analysis module is used to obtain statistical characteristic data of each state that becomes the state with the maximum forward metric by simulating the target CCSDS131.0-B-3 standard Turbo code trellis during the initialization or design phase. A baseline state register module is used to store baseline state identifiers selected and determined based on the statistical characteristic data; During real-time decoding, the forward recursive calculation and normalization module subtracts the current metric value of the base state from the path metrics of all states in parallel within each clock cycle, based on the state identifier stored in the base state register module, so as to synchronously complete the forward metric recursive calculation and normalization operation within one clock cycle. The interleaving / deinterleaving unit is used to interleave or deinterleave the soft information at the input and output of the component decoder; The control and interface unit is used to enable the forward metric normalization system for Turbo decoding to interact with external systems in terms of data and control signals.

[0013] Preferably, the forward metric normalization system for Turbo decoding of the CCSDS131.0-B-3 standard is integrated into a sliding window pipeline architecture; The sliding window pipeline architecture includes multiple sequentially connected LOG-MAP processing stages. Each LOG-MAP processing stage is responsible for processing a sliding window of decoded frame data and uses ping-pong buffer operations to hide the initialization delay of backward metrics, enabling the forward recursive computation and normalization module to complete all operations within a single clock cycle.

[0014] Based on the same concept, the present invention also provides an electronic device including a processor and a memory, the memory storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard as described above.

[0015] Based on the same concept, the present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions, which, when called and executed by a processor, cause the processor to implement the forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard as described above.

[0016] Because the present invention adopts the above technical solution, it has the following advantages and positive effects compared with the prior art: This invention provides a forward metric normalization method, system, electronic device, and storage medium for CCSDS131.0-B-3 standard Turbo decoding. First, a systematic simulation analysis is performed on the numerical change behavior of the forward metric Alpha for each state during the decoding process of the CCSDS131.0-B-3 standard Turbo code frame. The statistical characteristics of each state becoming the state with the maximum forward metric Alpha during the forward recursion process are obtained. Then, based on these statistical characteristics, a fixed state is selected from all states of the CCSDS131.0-B-3 standard Turbo code trellis as the base state for forward metric Alpha normalization throughout the entire decoding process. Finally, during the recursive calculation of forward metric Alpha using the Log-MAP algorithm, for each time t in the decoding process, the unnormalized forward metric values ​​of all states at the current time are subtracted from the metric value of the base state at the current time t, thereby completing the normalization operation of the forward metric Alpha for all states at that time. This invention simplifies normalization from a dynamic comparison of O(N) to a fixed subtraction of O(1), greatly reducing hardware resource consumption and critical path latency. At the same time, it enables the complete LOG-MAP metric recursion (including normalization) to be completed in a single clock cycle, perfectly adapting to high-throughput pipeline architectures. Furthermore, the baseline state selected based on statistical characteristics can effectively maintain the numerical stability of the metric, ensuring decoding gain and reducing hardware resource consumption. Attached Figure Description

[0017] Figure 1 This invention provides a flowchart illustrating a forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard; Figure 2 The grid state transition diagram of the CCSDS131.0-B-3 standard Turbo code encoder provided by this invention; Figure 3 The present invention provides a probability and interval distribution diagram of each state becoming the maximum value of Alpha metric under a regular input sequence; Figure 4The probability and interval distribution diagram of each state becoming the maximum value of the Alpha metric under a random input sequence provided by this invention. Figure 5 Timing diagram of the LOG-MAP sliding window pipeline architecture provided by this invention; Figure 6 A schematic diagram of the implementation of the LOG-MAP pipeline architecture provided by this invention in Simulink HDL; Figure 7 The actual delay measurement results of the LOG-MAP processing procedure provided by this invention are shown in the figure. Figure 8 The present invention provides a complete schematic diagram of the CCSDS131.0-B-3 standard Turbo iterative decoding process; Figure 9 The overall architecture of the CCSDS131.0-B-3 standard Turbo decoder provided by this invention is illustrated in the Simulink HDL implementation diagram. Figure 10 Resource usage analysis diagram of the decoder provided by this invention on a Xilinx Kintex-7 410T FPGA; Figure 11 The timing performance (clock frequency margin) analysis diagram of the decoder provided by this invention; Figure 12 The present invention provides an analysis diagram of the error rate performance (decoding gain) of the decoder at different code rates. Detailed Implementation

[0018] The following detailed description, in conjunction with the accompanying drawings and specific embodiments, provides a forward metric normalization method, system, electronic device, and storage medium for Turbo decoding of the CCSDS131.0-B-3 standard proposed in this invention. The advantages and features of this invention will become clearer from the following description and claims.

[0019] First Embodiment See Figure 1 This embodiment provides a forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding, which prevents numerical overflow when implementing LOG-MAP-like algorithms in CCSDS131.0-B-3 standard Turbo decoding. The specific steps of the forward metric normalization method include: S1: Conduct a systematic simulation analysis of the numerical change behavior of the forward metric Alpha in each state during the decoding process of the CCSDS131.0-B-3 standard Turbo code frame, and obtain the statistical characteristics of each state becoming the state with the maximum value of the forward metric Alpha during the forward recursion process.

[0020] Specifically, a systematic and large-scale Monte Carlo simulation analysis was conducted on the numerical change behavior of the forward metric Alpha of each state during the decoding process of the CCSDS131.0-B-3 standard Turbo code within the frame length defined by the standard (limited to 8920 bits in this embodiment). The simulation covered deterministic input sequences (such as alternating "0101..." codes) and random input sequence scenarios to ensure the robustness of the statistical results. Through simulation, the statistical characteristics of each state becoming the state with the maximum forward metric Alpha during the forward recursion process were obtained.

[0021] In the deterministic input mode, a periodic sequence such as "010101..." is input to test the measurement behavior of the grid state under extremely regular signals; in the random input mode, a randomly generated binary sequence is input to simulate a real communication environment.

[0022] S2: Based on statistical properties, select a fixed state from all states of the CCSDS131.0-B-3 standard Turbo code trellis diagram as the reference state for normalizing the forward metric Alpha throughout the entire decoding process.

[0023] Specifically, based on the detailed statistical characteristics obtained in step S1, an optimal fixed state is selected from all states of the CCSDS131.0-B-3 standard Turbo code trellis diagram according to scientific criteria. This fixed state serves as the reference state for the normalization of the forward metric Alpha throughout the decoding process. This selection is offline and one-time, and once selected, it is permanently incorporated into the decoder design.

[0024] S3: When the Log-MAP algorithm recursively calculates the forward metric Alpha, for each time t in the decoding process, the unnormalized forward metric values ​​of all states at the current time are subtracted from the metric value of the base state at the current time t, thereby completing the normalization operation of the forward metric Alpha of all states at that time.

[0025] Specifically, during the real-time recursive calculation of the forward metric Alpha based on the Log-MAP algorithm, for each time t in the decoding process, the unnormalized forward metric values ​​of all states at the current time are subtracted in parallel and uniformly from the metric value of the fixed reference state at the current time t, as determined in step S2. This subtraction operation is directly embedded in the forward recursive calculation formula and is completed synchronously with operations such as the accumulation of branch metrics and the comparison and selection between states within one clock cycle.

[0026] The following will provide a more detailed explanation of the specific implementation steps and functions of the forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding provided in this embodiment: Preferably, in one embodiment, the statistical features described in step S1 include a first statistic and a second statistic.

[0027] The first statistic is configured as the probability P_max(s) of each state of the CCSDS131.0-B-3 standard Turbo code trellis graph becoming the state with the forward metric alpha maximum during the decoding process of a complete frame of data, where Pmax(s) = Nmax(s) / L, and s represents the state index and L is the frame length; the second statistic is configured as the statistical distribution characteristic of the time interval Δt between two consecutive times that a state becomes the state with the forward metric alpha maximum for each state.

[0028] Furthermore, in one embodiment, the second statistic includes the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval Δt.

[0029] In step S2, a fixed state is selected as the reference state for the forward metric Alpha normalization throughout the entire decoding process. The selection criteria for the reference state are as follows: Select a state s0, and limit the probability P_max(s0) of s0 becoming the maximum value state to be higher than a preset first threshold, and the standard deviation σ_Δt(s0) of the time interval Δt between s0 becoming the maximum value state to be lower than a preset second threshold.

[0030] Furthermore, in one embodiment, selecting a fixed state in step S2 as the reference state for forward metric Alpha normalization throughout the decoding process includes the following steps: S21: As Figure 2 As shown, for the Turbo code trellis diagram of CCSDS131.0-B-3 standard, Monte Carlo simulations are performed with deterministic test sequences and randomized input sequences respectively, under the frame length defined by the standard. Calculate the probability P_max(s) of each state becoming the maximum state of the forward metric Alpha under the two input sequences, and calculate the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval sequence in which each state becomes the maximum state; The state that exhibits frequency meeting the first threshold requirement and time stability meeting the second threshold requirement under both input sequences is selected as the baseline state.

[0031] Specifically, the statistical characteristics described in step S1 mainly include two key indicators: first, the probability P_max(s) of each state becoming the maximum value state of the forward metric; second, the statistical distribution characteristics of the time interval Δt between two consecutive maximum values ​​of each state, especially its mean μ_Δt(s) and standard deviation σ_Δt(s). An ideal benchmark state needs to have the characteristics of "high frequency" and "high stability", that is, a high probability of becoming the maximum value (ensuring the representativeness of the normalized benchmark), and small fluctuations and low standard deviations in the time interval between becoming the maximum value (ensuring the timeliness and uniformity of the normalization operation and avoiding numerical risks caused by a long period without effective normalization).

[0032] like Figure 3 and Figure 4 As shown, in this embodiment, extensive simulation analysis of the CCSDS131.0-B-3 standard Turbo code trellis revealed that states 4 and 13 (corresponding to specific state numbers in the trellis) exhibit excellent statistical properties under various input conditions: their probability of becoming the maximum value state remains stable at approximately 6.5%, and the statistical distribution of their time intervals is highly concentrated, with over 95% probability falling within [0, 40] time units, and the standard deviation σ_Δt is relatively small (approximately 12). Therefore, in this embodiment, state 4 or state 13 is preferably a fixed normalized reference state suitable for the CCSDS131.0-B-3 standard Turbo code.

[0033] Preferably, in one embodiment, after the reference state is selected, it is fixed and applied to the Turbo decoding task of the CCSDS131.0-B-3 standard.

[0034] In this embodiment, the selection of the fixed reference state in step S2 is completed once. That is, after the simulation analysis of step S1 is completed and the reference state is selected for the CCSDS131.0-B-3 standard Turbo code, the reference state is fixed and applied to all decoding tasks that conform to the standard, without the need to select it again each time decoding.

[0035] Preferably, in one embodiment, the Log-MAP algorithm is used in S3 to recursively calculate the forward metric Alpha. The mathematical expression for the recursive calculation is: A t+1 (v) = max(A) t (u) - A t (s0) + Γ t (u, v) Where u is a grid state at time t, v is a valid grid state connected to u at time t+1, and (u,v) represents a valid state transition relationship. A t(u) is the normalized forward metric of the state u at time t, A t (s0) represents the normalized forward metric of the baseline state at time t, Γ t (u, v) is the branch metric for the transition from state u to state v at time t.

[0036] Compared to the traditional forward recursive formula that does not include normalization: A t+1 (v) = max(A) t (u) + Γ t (u,v)), where the max operation iterates through all possible states u that could transition to state v at the previous moment. In this embodiment, based on A t+1 (v) = max(A) t (u) - A t (s0) + Γ t The improved algorithm (u, v) allows for a subtraction operation that can be implemented in hardware with the branch metric Γ. t The addition operations of (u, v) are combined and used as part of the "path metric generation" logic, followed by the comparison and selection logic. Since the subtracted value is a metric of a certain state, rather than a dynamically searched maximum value, a complex comparator tree is not required, greatly simplifying the critical path.

[0037] Second Embodiment Based on the same concept, the present invention also provides a forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding, for performing the forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in any one of the first embodiments, comprising: The statistical characteristic analysis module is used to obtain statistical characteristic data of each state that becomes the state with the maximum forward metric by simulating the target CCSDS131.0-B-3 standard Turbo code trellis during the initialization or design phase. The baseline state register module is used to store the baseline state identifier selected and determined based on statistical characteristic data; During real-time decoding, the forward recursive computation and normalization module subtracts the current metric value of the base state from the path metrics of all states in parallel within each clock cycle, based on the state identifier stored in the base state register module, so as to synchronously complete the forward metric recursive computation and normalization operation within one clock cycle. The interleaving / deinterleaving unit is used to interleave or deinterleave the soft information at the input and output of the component decoder; The control and interface unit is used to enable the forward metric normalization system for Turbo decoding to interact with external systems in terms of data and control signals.

[0038] In this embodiment, a complete CCSDS131.0-B-3 standard Turbo decoder is constructed based on a statistical characteristic analysis module, a reference state register module, a forward recursive calculation and normalization module, an interleaving / deinterleaving unit, and a control and interface unit. Figure 8 and Figure 9 As shown.

[0039] Preferably, in one embodiment, the forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding is integrated into a sliding window pipeline architecture.

[0040] The sliding window pipeline architecture consists of multiple sequentially connected LOG-MAP processing stages. Each LOG-MAP processing stage is responsible for processing a sliding window of decoded frame data and uses ping-pong buffer operations to hide the initialization delay of backward metrics, enabling the forward recursive computation and normalization modules to complete all operations within a single clock cycle.

[0041] In this embodiment, to achieve high throughput, the following is adopted: Figure 5 The sliding window pipeline architecture is shown. This architecture divides a long frame of data into several consecutive, partially overlapping windows. Multiple identical LOG-MAP processing cores operate in a deep pipeline manner, with each core handling the forward and backward recursion of one window. Since backward recursion requires initialization from the end of the window, resulting in "initialization bubbles," the sliding window architecture cleverly overlaps the forward recursion of the next window with the backward recursion of the current window in time through a double-buffer mechanism using "ping-pong" operations, hiding the initialization delay and improving hardware utilization. Because the normalization operation described in this embodiment is simplified to a single subtraction, it makes... Figure 6 The forward recursive computation unit (Alpha Unit) inside the LOG-MAP core, as shown, can complete the entire process within one clock cycle, from reading the previous time step metric, adding the branch metric, subtracting the baseline state metric, to comparing and writing the new metric. For example... Figure 7 Actual delay measurements show that the longest combinational logic path delay is much less than one clock cycle (e.g., 16.28 ns at 61.44 MHz, while the critical path delay is only about 13 ns), thus providing ample setup time margin (e.g., ...). Figure 11 As shown, there is still a 3ns margin under a 90MHz constraint.

[0042] Furthermore, performance verification of a forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding on a Xilinx Kintex-7 410T FPGA platform shows that, as Figure 10 As shown, the decoder consumes only about 6% of the FPGA's lookup table (LUT) and register resources, demonstrating the significant advantage of this system and decoder in reducing hardware complexity. Furthermore, static timing analysis shows that at the target operating frequency of 61.44MHz, there is sufficient timing margin, and even if the constraint frequency is increased to 90MHz, the timing requirements can still be met. Figure 10 This demonstrates its superior timing characteristics, leaving room for higher-speed communication requirements. Furthermore, as... Figure 12 As shown, significant decoding gains (5~8.5 dB) can be obtained at different code rates, fully meeting the real-time decoding requirements of CCSDS low-orbit (512fps) and deep space (32fps) scenarios.

[0043] This embodiment also provides an electronic device, including a processor and a memory. The memory stores machine-executable instructions that can be executed by the processor. The processor executes the machine-executable instructions to implement the aforementioned forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard.

[0044] This embodiment also provides a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions cause the processor to implement the aforementioned forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard.

[0045] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, if these changes fall within the scope of the claims of the present invention and their equivalents, they shall still fall within the protection scope of the present invention.

Claims

1. A forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding, characterized in that, Includes the following steps: S1: Conduct a systematic simulation analysis of the numerical change behavior of the forward metric Alpha in each state during the decoding process of the CCSDS131.0-B-3 standard Turbo code frame, and obtain the statistical characteristics of each state becoming the state with the maximum value of the forward metric Alpha during the forward recursion process. S2: Based on the statistical characteristics, select a fixed state from all states of the CCSDS131.0-B-3 standard Turbo code trellis diagram as the reference state for forward metric Alpha normalization in the entire decoding process; S3: When the Log-MAP algorithm performs recursive calculation of the forward metric Alpha, for each time t in the decoding process, the unnormalized forward metric values ​​of all states at the current time are subtracted from the metric value of the base state at the current time t, thereby completing the normalization operation of the forward metric Alpha of all states at that time.

2. The forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in claim 1, characterized in that, The statistical characteristics include a first statistic and a second statistic; The first statistic is configured as the probability P_max(s) of each state of the CCSDS131.0-B-3 standard Turbo code trellis graph becoming the state with the forward metric alpha maximum during the decoding of a complete frame of data, where s represents the state index; The second statistic is configured to be the statistical distribution characteristic of the time interval Δt experienced between two consecutive states that are the maximum values ​​of the forward metric Alpha for each state.

3. The forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in claim 2, characterized in that, The second statistic includes the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval Δt; In step S2, a fixed state is selected as the reference state for the forward metric Alpha normalization throughout the entire decoding process. The selection criteria for the reference state are as follows: Select a state s0, and limit the probability P_max(s0) of s0 becoming the maximum value state to be higher than a preset first threshold, and the standard deviation σ_Δt(s0) of the time interval Δt between s0 becoming the maximum value state to be lower than a preset second threshold.

4. The forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in claim 3, characterized in that, In step S2, a fixed state is selected as the reference state for the forward metric Alpha normalization throughout the entire decoding process, including the following steps: S21: For the Turbo code trellis diagram of CCSDS131.0-B-3 standard, Monte Carlo simulation is performed by inputting deterministic test sequences and randomized input sequences under the frame length defined in the standard. Calculate the probability P_max(s) of each state becoming the maximum state of the forward metric Alpha under the two input sequences, and calculate the mean μ_Δt(s) and standard deviation σ_Δt(s) of the time interval sequence in which each state becomes the maximum state; The state that exhibits both probability frequency and time stability that meet the first threshold requirement under both input sequences is selected as the baseline state. The reference state is defined as state 4 or state 13 in the 16-group state coding of the CCSDS131.0-B-3 standard Turbo code trellis diagram.

5. The forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard as described in any one of claims 1-4, characterized in that, Once the reference state selection is completed, it is fixed and applied to the decoding task of the CCSDS131.0-B-3 standard.

6. The forward metric normalization method for CCSDS131.0-B-3 standard Turbo decoding as described in claim 1, characterized in that, In S3, the Log-MAP algorithm recursively calculates the forward metric Alpha. The mathematical expression for the recursive calculation is: A t+1 (v) = max( A t (u) - A t (s0) + Γ t (u, v) ) Where u is a grid state at time t, v is a valid grid state connected to u at time t+1, and (u,v) represents a valid state transition relationship. A t (u) is the normalized forward metric of the state u at time t, A t (s0) is the normalized forward metric value of the reference state at time t, Γ t (u, v) is the branch metric for the transition from state u to state v at time t.

7. A forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding, characterized in that, A forward metric normalization method for performing Turbo decoding of the CCSDS131.0-B-3 standard as described in any one of claims 1-6, comprising: The statistical characteristic analysis module is used to obtain statistical characteristic data of each state that becomes the state with the maximum forward metric by simulating the target CCSDS131.0-B-3 standard Turbo code trellis during the initialization or design phase. A baseline state register module is used to store baseline state identifiers selected and determined based on the statistical characteristic data; During real-time decoding, the forward recursive calculation and normalization module subtracts the current metric value of the base state from the path metrics of all states in parallel within each clock cycle, based on the state identifier stored in the base state register module, so as to synchronously complete the forward metric recursive calculation and normalization operation within one clock cycle. The interleaving / deinterleaving unit is used to interleave or deinterleave the soft information at the input and output of the component decoder; The control and interface unit is used to enable the forward metric normalization system for Turbo decoding to interact with external systems in terms of data and control signals.

8. The forward metric normalization system for CCSDS131.0-B-3 standard Turbo decoding as described in claim 7, characterized in that, Integrated into the sliding window pipeline architecture; The sliding window pipeline architecture includes multiple sequentially connected LOG-MAP processing stages. Each LOG-MAP processing stage is responsible for processing a sliding window of decoded frame data and uses ping-pong buffer operations to hide the initialization delay of backward metrics, enabling the forward recursive computation and normalization module to complete all operations within a single clock cycle.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing machine-executable instructions that can be executed by the processor, the processor executing the machine-executable instructions to implement the forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the forward metric normalization method for Turbo decoding of the CCSDS131.0-B-3 standard as described in any one of claims 1-6.

Images