Self-adaptive error correction method in nuclear signal digital coincidence measurement

Abstract

The invention provides an adaptive error correction method in nuclear signal digital coincidence measurement, and the method comprises the steps: obtaining an original data stream in the nuclear signal digital coincidence measurement, extracting the dynamic range and time-varying characteristic parameters of a signal, and building a signal feature model; according to the signal feature model, the working condition type of the signal is judged, and the error correction demand level under the current working condition is determined; acquiring hardware condition parameters of the measurement system, including storage capacity and computing power, and evaluating the feasibility of the selected error correction strategy; in the data processing process, monitoring signal feature changes in real time, and if it is detected that the signal features exceed a preset threshold value, re-evaluating the working condition type; and according to a re-evaluation result, dynamically adjusting an error correction strategy, and realizing adaptive balance between the error correction capability and the real-time performance. According to the invention, the accuracy and reliability of nuclear signal measurement are obviously improved, and an important guarantee is provided for safe operation of nuclear facilities.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an adaptive error correction method in the digital coincidence measurement of nuclear signals. Background Technology

[0002] Background of the problem: In nuclear signal digital coincidence measurement, digital error correction technology faces a key technical dilemma. On the one hand, to ensure the accuracy and reliability of the measurement results, comprehensive error detection and correction of the acquired digital signal data are required. This necessitates error correction algorithms with high error correction capabilities, capable of detecting and correcting various types of errors, such as random errors and burst errors. However, on the other hand, nuclear signal measurement demands high real-time performance, requiring data acquisition, transmission, and processing to be completed within an extremely short timeframe. Overly complex error correction algorithms significantly increase computational load and processing latency, impacting the real-time performance of the measurement.

[0003] Furthermore, nuclear signals have a wide dynamic range and strong time-varying characteristics, with significant differences in signal characteristics under different measurement scenarios. A uniform error correction strategy is difficult to achieve optimal performance under various operating conditions. How to adaptively adjust the error correction strategy according to signal characteristics and measurement requirements, ensuring both measurement accuracy and real-time performance, is a pressing technical challenge. Simultaneously, the error correction algorithm must also consider the hardware limitations of the measurement system, such as storage capacity and computing power. Achieving optimal error correction performance with limited resources places higher demands on algorithm design and optimization. Summary of the Invention

[0004] This invention provides an adaptive error correction method for nuclear signal digital coincidence measurement, the method comprising: Obtain the raw data stream from the digital coincidence measurement of the nuclear signal, extract the dynamic range and time-varying characteristic parameters of the signal, and establish a signal characteristic model; Based on the signal characteristic model, determine the type of operating condition the signal is in and determine the level of error correction requirement under the current operating condition; A multi-level error correction strategy library is pre-established, with each level of strategy corresponding to different error correction capabilities and computational complexity. The strategy library contains error correction algorithms for random errors and burst errors. Based on the current operating condition type and the level of error correction requirement, the optimal error correction strategy is matched from the strategy library to determine the type of error correction algorithm and parameter configuration to be used. During data processing, changes in signal characteristics are monitored in real time. If a signal characteristic is detected to exceed a preset threshold, the operating condition type is reassessed. Based on the reassessment results, the error correction strategy is dynamically adjusted to achieve an adaptive balance between error correction capability and real-time performance.

[0005] Preferably, the process of acquiring the raw data stream from the nuclear signal digital conformation measurement, extracting the dynamic range and time-varying characteristic parameters of the signal, and establishing a signal feature model includes: Acquire the raw data stream of the nuclear signal, preprocess the data, and remove noise and interference. Based on the time-domain and frequency-domain characteristics of the nuclear signal, the dynamic range and time-varying characteristic parameters of the signal are extracted; Feature selection is performed on the extracted characteristic parameters to identify key parameters that contribute to signal feature modeling; Machine learning algorithms, such as support vector machines or neural networks, are used to construct a feature model of nuclear signals; The performance of the established signal feature model is evaluated and optimized using cross-validation. The optimized nuclear signal feature model was applied to real data to assess its generalization ability and robustness. Based on the output of the nuclear signal characteristic model, the dynamic range and time-varying characteristics of the nuclear signal are determined, providing a basis for subsequent signal analysis.

[0006] Preferably, the step of determining the operating condition type of the signal based on the signal characteristic model and determining the error correction requirement level under the current operating condition includes: By extracting features from the signal, the feature vector of the signal is obtained, which serves as the basis for determining the type of operating condition of the signal. Based on the pre-established signal feature model, the extracted signal feature vector is input into the model, and the operating condition type of the signal is obtained through model reasoning. For different working conditions, different levels of error correction requirements are preset, and the corresponding error correction requirement level is obtained based on the working condition type determined by the model. If the matching degree between the feature vector of the signal and the feature model of any working condition type is lower than the preset threshold, it is determined to be an unknown working condition and the default error correction strategy is adopted. Based on the determined error correction requirement level, select the corresponding error correction strategy from the preset error correction strategy library and apply it to the error correction processing of the signal. During the training process of the signal feature model, the support vector machine algorithm is used to model the signal features under different working conditions, and the feature discrimination model for each working condition is obtained. When classifying error correction requirement levels, the signal quality requirements and fault tolerance capabilities under different operating conditions are comprehensively considered, and different error correction requirement levels and corresponding error correction strategies are set accordingly.

[0007] Preferably, the pre-established multi-level error correction strategy library, where each level corresponds to different error correction capabilities and computational complexity, includes error correction algorithms for random and burst errors, including: Based on different requirements for error correction capability and computational complexity, a corresponding multi-level error correction strategy library is established. Each level of the strategy library contains error correction algorithms for random errors and burst errors. Obtain the error type of the data to be corrected, determine whether it is a random error or a burst error, and select a suitable error correction algorithm from the corresponding error correction strategy library according to the error type. For different application scenarios and error correction requirements, select an error correction strategy from a multi-level error correction strategy library that meets the current error correction capability and computational complexity requirements; The data to be corrected is input into the selected error correction algorithm for processing. The error correction algorithm corrects the bit errors in the data to obtain the corrected data. The correctness of the corrected data is evaluated by comparing it with the original correct data and calculating the error correction rate of the error correction algorithm to verify the effectiveness of the selected error correction strategy. If the error correction rate does not meet expectations, a higher-level error correction strategy is selected from the multi-level error correction strategy library based on the trade-off between error correction rate and computational complexity, and the data is reprocessed for error correction. The above error correction process is iterated continuously until the corrected data meets the error correction requirements of the application scenario, and the final error correction result data is output.

[0008] Preferably, the step of matching the optimal error correction strategy from the strategy library based on the current operating condition type and error correction requirement level, and determining the type of error correction algorithm and parameter configuration to be used, includes: Obtain information on the current operating condition type and error correction requirement level as a basis for matching the optimal error correction strategy; In the preset strategy library, one or more error correction strategies are matched according to the current working condition type and the error correction requirement level. If multiple error correction strategies are matched, they are scored and ranked according to the preset evaluation rules, and the error correction strategy with the highest score is determined as the optimal error correction strategy. From the optimal error correction strategy, obtain information such as the error correction algorithm type and algorithm parameter configuration; Based on the type of error correction algorithm, the corresponding error correction algorithm model is called from the algorithm model library; Input the algorithm parameter configuration information into the error correction algorithm model to configure the parameters of the error correction algorithm model; Using a parameter-configured error correction algorithm model, the input data to be corrected is processed to obtain the corrected data output.

[0009] Preferably, after determining the type of error correction algorithm and parameter configuration, the hardware condition parameters of the measurement system are obtained, including storage capacity and computing power, and the feasibility of the selected error correction strategy is evaluated, including: Obtain the hardware parameters of the measurement system, including indicators such as storage capacity and computing power, and use them as the basis for evaluating the feasibility of the error correction strategy; Based on the storage capacity of the measurement system, determine the amount of data that the system can store, and select an appropriate error correction strategy accordingly. By analyzing the computational capabilities of the measurement system, it can be determined whether the system meets the computational complexity requirements of the selected error correction strategy. If not, the error correction strategy needs to be adjusted. The decision tree algorithm is used to automatically recommend the optimal error correction strategy by comprehensively considering the storage capacity and computing power of the measurement system. Obtain the hardware configuration information of the measurement system, divide it into different performance levels using a clustering algorithm, and pre-set an appropriate error correction strategy for each level; Establish a mapping relationship between measurement system hardware parameters and error correction strategies, and automatically adjust the corresponding error correction strategies when hardware parameters change; By using the support vector machine algorithm, an evaluation model for the matching degree between the hardware parameters of the measurement system and the error correction strategy is trained based on historical data, which can be used to guide the selection of the error correction strategy.

[0010] Preferably, after evaluating the feasibility of the selected error correction strategy, if the hardware conditions meet the strategy requirements, the corresponding error correction algorithm is loaded; if the hardware conditions are insufficient, the algorithm parameters are adjusted according to priority to reduce computational complexity, including: Obtain performance metrics of the current hardware conditions, including parameters such as memory capacity, processor frequency, and storage space, to form a hardware condition vector; Read the hardware requirements of each error correction algorithm from the policy library to form a policy requirement vector; The matching degree between the hardware condition vector and the policy requirement vector is calculated to obtain the fit score of each error correction algorithm. The error correction algorithms are prioritized based on their fitness scores. Algorithms that meet the hardware requirements have higher priority, while those that do not have lower priority. If the hardware conditions meet the requirements of the highest priority error correction algorithm, then load the algorithm directly and proceed to step 7; Otherwise, proceed to step 6; For error correction algorithms that do not meet the hardware requirements, their algorithm parameters are traversed from high to low priority. The parameters are adjusted by reducing data precision and decreasing the number of iterations until the hardware requirements are met, thus obtaining the corrected algorithm parameter set. Using a defined error correction algorithm and its parameter set, the received data is processed to correct errors, and the corrected data is output.

[0011] Preferably, during data processing, real-time monitoring of signal characteristic changes is performed. If a signal characteristic exceeds a preset threshold, the operating condition type is reassessed, including: Acquire real-time signal data during data processing and extract the characteristic parameters of the signal; The extracted signal feature parameters are compared with a preset threshold to determine whether they exceed the threshold range. If the signal characteristic parameters exceed the preset threshold, the re-evaluation process of the operating condition type will be triggered. Based on the trend of signal characteristic changes and historical operating condition types, the operating condition type is re-evaluated using the support vector machine algorithm; By using the decision tree algorithm, combined with the characteristics of signal feature changes and the type of operating condition being evaluated, it is determined whether the data processing flow needs to be adjusted. If the data processing workflow needs to be adjusted, the parameters of the data processing algorithm will be dynamically optimized according to the type of working condition being evaluated. The optimized data processing algorithm is applied to real-time signal data to obtain processing results adapted to the current working conditions, thus completing the dynamic adjustment of the data processing flow.

[0012] Preferably, the step of dynamically adjusting the error correction strategy based on the re-evaluation results to achieve an adaptive balance between error correction capability and real-time performance includes: Obtain the initial parameters and strategy of the error correction algorithm, perform error correction processing on the text data to be corrected, and obtain the initial error correction result; Based on the preset evaluation indicators, the initial error correction results are evaluated in real time, and indicators such as error correction accuracy and error correction speed are calculated to determine whether the preset performance threshold is met. If the real-time evaluation results do not meet the preset threshold, the parameters of the error correction algorithm are dynamically adjusted through machine learning algorithms based on the evaluation results, such as adjusting the error correction granularity, error correction rules and feature weights, to generate a new error correction strategy. The dynamically adjusted error correction strategy is applied to the error correction algorithm to reprocess the text data and obtain new error correction results. The new error correction results are evaluated in real time, relevant performance indicators are calculated, and it is determined whether the adjusted error correction strategy has effectively improved the error correction capability and real-time performance. If the new error correction strategy is effective, it is saved as the optimal error correction strategy and applied to subsequent error correction tasks. If ineffective, return to the pre-established multi-level error correction strategy library. Each level of strategy corresponds to different error correction capabilities and computational complexity. The strategy library contains error correction algorithms for random errors and burst errors, and the error correction strategy is adjusted again. Through continuous iteration and dynamic adjustments and evaluations, the error correction algorithm achieves an adaptive balance between its error correction capability and real-time performance, meeting business needs.

[0013] Preferably, after the error correction strategy is dynamically adjusted, the error-corrected measurement data is output, and a strategy adjustment log is recorded for subsequent strategy optimization, including: Acquire measurement data, perform error correction processing on the measurement data according to preset error correction rules, and obtain the error-corrected measurement data; Determine whether the policy adjustment conditions are triggered during the error correction process. If so, record the policy adjustment log, including the adjusted policy parameters and the reason for the adjustment. Output the corrected measurement data and save the strategy adjustment log as a reference for subsequent optimization of the error correction strategy. The optimized error correction strategy model is obtained by training historical measurement data and corresponding error correction results using machine learning algorithms. The optimized error correction strategy model is applied to new measurement data to perform error correction and the error correction effect is evaluated. Based on the evaluation results of the error correction effect, determine whether further optimization of the error correction strategy is needed; The final optimized error correction strategy model is solidified and applied to subsequent measurement data error correction processing. The error correction effect is continuously monitored, and a new round of strategy optimization is initiated when necessary.

[0014] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses an adaptive error correction method for digital coincidence measurement of nuclear signals. The method analyzes the dynamic range and time-varying characteristics of the raw data stream to establish a signal characteristic model, and then determines the current operating condition type and error correction requirement level. This invention pre-establishes a multi-level error correction strategy library, automatically selecting the optimal error correction strategy based on the operating condition type and hardware conditions. During data processing, this invention can monitor changes in signal characteristics in real time and dynamically adjust the error correction strategy, achieving an adaptive balance between error correction capability and real-time performance. This method can effectively handle random and sudden bit errors, and can flexibly adjust algorithm parameters according to hardware conditions, ensuring optimal error correction results under different measurement environments. By dynamically optimizing the error correction strategy, this invention significantly improves the accuracy and reliability of nuclear signal measurement, providing an important guarantee for the safe operation of nuclear facilities. Attached Figure Description

[0015] Figure 1 This is a flowchart of an adaptive error correction method in nuclear signal digital coincidence measurement according to the present invention. Detailed Implementation

[0016] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0017] like Figure 1 This embodiment of an adaptive error correction method in nuclear signal digital coincidence measurement may specifically include: Step S101: Obtain the raw data stream from the nuclear signal digital coincidence measurement, extract the dynamic range and time-varying characteristic parameters of the signal, and establish a signal characteristic model.

[0018] The process involves acquiring the raw nuclear signal data stream, preprocessing the data to remove noise and interference, and extracting dynamic range and time-varying characteristic parameters based on the time and frequency domain characteristics of the nuclear signal. Feature selection is then performed on the extracted parameters to identify key parameters that contribute to signal feature modeling. Machine learning algorithms, such as support vector machines or neural networks, are used to construct a nuclear signal feature model. Cross-validation is employed to evaluate and optimize the performance of the established signal feature model. The optimized nuclear signal feature model is then applied to real-world data to assess its generalization ability and robustness. Based on the output of the nuclear signal feature model, the dynamic range and time-varying characteristics of the nuclear signal are determined, providing a basis for subsequent signal analysis.

[0019] Specifically, firstly, the raw data stream of the nuclear signal was acquired using high-precision data acquisition equipment, with a sampling frequency of 10kHz and a sampling precision of 16 bits. Then, the acquired raw data was preprocessed, employing wavelet transform for noise reduction. The wavelet basis function was selected as sym4, and the decomposition level was 5, removing high-frequency noise and interference information from the signal. Next, based on the time-domain and frequency-domain characteristics of the nuclear signal, the dynamic range and time-varying characteristic parameters of the signal were extracted, including time-domain parameters such as peak-to-peak value, root mean square value, kurtosis, and skewness, and frequency-domain parameters such as center frequency, bandwidth, and spectral entropy, resulting in a total of 20 characteristic parameters. Feature selection was performed on the extracted characteristic parameters, using the minimum redundancy maximum correlation algorithm to select 10 key parameters that contribute to signal feature modeling. Then, a support vector machine algorithm was used to construct a nuclear signal feature model, with the radial basis function chosen as the kernel function, a penalty factor C of 10, and 1000 training samples. The performance of the established signal feature model was evaluated using a 5-fold cross-validation method, achieving an average recognition accuracy of over 95%. The optimized nuclear signal feature model was applied to another 500 real-world data samples to assess its generalization ability and robustness. The results showed that the model can effectively extract and identify features of nuclear signals under different operating conditions. Finally, based on the output of the nuclear signal feature model, the dynamic range of the nuclear signal was determined to be -5V to +5V, with the main frequency components concentrated between 1kHz and 3kHz, providing an important basis for subsequent signal analysis and fault diagnosis.

[0020] Step S102: Based on the signal characteristic model, determine the type of operating condition of the signal and determine the error correction requirement level under the current operating condition.

[0021] Feature extraction is performed on the signal to obtain its feature vector, which serves as the basis for determining the signal's operating condition type. Based on a pre-established signal feature model, the extracted feature vector is input into the model, and the model infers the operating condition type to which the signal belongs. Different error correction requirements are preset for different operating condition types, and the corresponding error correction requirement level is obtained based on the operating condition type determined by the model. If the matching degree between the signal's feature vector and the feature model of any operating condition type is lower than a preset threshold, it is determined to be an unknown operating condition, and the default error correction strategy is adopted. Based on the determined error correction requirement level, an error correction strategy of the corresponding level is selected from a preset error correction strategy library and applied to the error correction processing of the signal. During the training process of the signal feature model, a support vector machine algorithm is used to model the signal features under different operating conditions, obtaining feature discrimination models for each operating condition. When classifying the error correction requirement level, the quality requirements and fault tolerance capabilities of the signal under different operating conditions are comprehensively considered, and different error correction requirement levels and corresponding error correction strategies are set accordingly.

[0022] Specifically, in the signal feature extraction process, wavelet transform is used to perform time-frequency domain analysis on the signal, extracting statistical features such as mean, variance, and skewness, as well as frequency domain features such as center frequency and bandwidth, forming a feature vector of dimension 20. Using 1000 labeled signal samples under different operating conditions, the SVM algorithm is used to train the signal features, and a condition discrimination model with an average recognition accuracy of 92% is obtained through 5-fold cross-validation. When determining the error correction requirement level, the highest level of error correction requirement is set for control command signals with high reliability requirements. High-performance error correction coding methods such as Turbo codes and LDPC codes are used, combined with CRC check and ARQ retransmission measures to ensure the reliability of information transmission. For data acquisition signals, different error correction levels are divided according to the importance of the signal, and relatively simple error correction strategies such as RS codes and convolutional codes are used to meet reliability requirements while also considering transmission efficiency. Upon receiving a new signal, wavelet transform is used to extract signal features, which are then input into the SVM model for operating condition discrimination. If the confidence level exceeds 8, a preset error correction strategy is indexed and applied to the signal based on the identified operating condition type. If the confidence level is below 8, the signal is classified as an unknown operating condition, and the default RS(255,239) code scheme is used for encoding and decoding. Simulation experiments demonstrate that this method can adaptively select appropriate error correction strategies based on the characteristics of different operating condition signals, maximizing transmission efficiency while ensuring transmission reliability. Compared to the traditional "one-size-fits-all" error correction scheme, performance is improved by more than 15%.

[0023] Step S103: A multi-level error correction strategy library is pre-established. Each level of strategy corresponds to different error correction capabilities and computational complexity. The strategy library contains error correction algorithms for random errors and burst errors.

[0024] Based on varying requirements for error correction capability and computational complexity, a multi-level error correction strategy library is established, with each level containing error correction algorithms for both random and burst errors. The error type of the data to be corrected is obtained, determining whether it is a random or burst error. A suitable error correction algorithm is selected from the corresponding strategy library based on the error type. For different application scenarios and error correction requirements, an error correction strategy that meets the current error correction capability and computational complexity requirements is selected from the multi-level strategy library. The data to be corrected is input into the selected error correction algorithm for processing. The algorithm corrects the errors in the data, resulting in corrected data. The correctness of the corrected data is evaluated by comparing it with the original correct data to calculate the error correction rate, thus verifying the effectiveness of the selected error correction strategy. If the error correction rate does not meet expectations, a higher-level error correction strategy is selected from the multi-level strategy library based on a trade-off between error correction rate and computational complexity, and the data is reprocessed for error correction. This error correction process is iterated until the corrected data meets the error correction requirements of the application scenario, at which point the final corrected data is output.

[0025] Specifically, when establishing a multi-level error correction strategy library, three levels can be set up according to different requirements for error correction capability and computational complexity. The first-level strategy library targets scenarios with low error correction capability requirements but high computational complexity requirements, employing simple error correction algorithms such as Hamming codes and cyclic redundancy check (CRC) codes. The second-level strategy library targets scenarios with moderate error correction capability and computational complexity requirements, employing medium-complexity error correction algorithms such as convolutional codes and Reed-Solomon codes. The third-level strategy library targets scenarios with high error correction capability requirements but low computational complexity requirements, employing high-complexity error correction algorithms such as LDPC codes and Turbo codes. When determining the type of bit error, the distribution characteristics of the bit error can be statistically analyzed. For example, random bit errors typically exhibit a uniform distribution; while burst bit errors are usually concentrated in a specific region of the data. Based on this characteristic, a bit error distribution statistical algorithm can be designed to calculate the bit error density in different regions of the data. When the bit error density in a certain region is significantly higher than in other regions, it is determined to be a burst bit error; otherwise, it is determined to be a random bit error. When selecting a suitable error correction strategy, the requirements of the current application scenario regarding error correction capability and computational complexity can be comprehensively considered. The optimal strategy can be selected from a multi-level strategy library using a weighted scoring method. For example, two scoring criteria can be set: error correction capability and computational complexity. Each error correction strategy can be scored, and then different weights can be assigned to the scores of the two dimensions according to the emphasis of the application scenario. Finally, the error correction strategy with the highest weighted score is selected. When evaluating error correction performance, the bit error rate (BER) metric can be used, which is the proportion of residual errors in the corrected data. By comparing the corrected data with the original correct data, the number of errors in the corrected data is counted, and the BER is calculated. For example, if there are still 5 errors in 1000 bits of corrected data, the BER is 5%. A threshold for the BER can be set according to the specific requirements of the application scenario. When the BER after error correction exceeds the threshold, the currently selected error correction strategy is considered insufficient, and a higher-level strategy needs to be selected for re-error correction.

[0026] Step S104: Based on the current working condition type and the error correction requirement level, match the optimal error correction strategy from the strategy library and determine the error correction algorithm type and parameter configuration to be used.

[0027] The system acquires the current operating condition type and error correction requirement level information as the basis for matching the optimal error correction strategy. From a pre-set strategy library, it matches one or more corresponding error correction strategies based on the current operating condition type and error correction requirement level. If multiple error correction strategies are matched, they are scored and ranked according to pre-set evaluation rules, and the strategy with the highest score is determined as the optimal error correction strategy. From the optimal error correction strategy, it acquires information such as the error correction algorithm type and algorithm parameter configuration. Based on the error correction algorithm type, it calls the corresponding error correction algorithm model from the algorithm model library. It inputs the algorithm parameter configuration information into the error correction algorithm model to configure its parameters. Using the parameter-configured error correction algorithm model, it performs error correction processing on the input data to be corrected, obtaining the corrected data output result.

[0028] Specifically, firstly, the system collects real-time environmental data, such as temperature, humidity, and vibration frequency, using sensors, and determines the current operating condition type based on preset thresholds. For example, when the ambient temperature is above 40℃ and the humidity is below 30%, it is determined to be a high-temperature dry operating condition; when the vibration frequency exceeds 50Hz, it is determined to be a high-frequency vibration operating condition. Simultaneously, users can set the error correction requirement level through the interface, such as high, medium, and low. The system uses the operating condition type and error correction requirement level as two dimensions to match the optimal error correction strategy in the strategy library. If multiple strategies are matched, they are scored and ranked according to preset evaluation rules, such as comprehensively considering factors like error correction efficiency, error correction accuracy, and resource consumption, assigning different weight coefficients to each strategy. The strategy with the highest score is selected as the optimal strategy, and its algorithm type and parameter configuration are extracted. For example, the optimal strategy might be "convolutional neural network," with parameters such as "learning rate 01, iteration count 500, regularization coefficient 02," etc. Then, the system calls the corresponding model from the algorithm model library based on the algorithm type, imports the parameter configuration into the model, and completes model initialization. Finally, the data to be corrected is input into the model for calculation and processing, and the model output is the corrected and clear data. Through this series of processes, intelligent error correction that adapts to operating conditions and matches requirements is achieved.

[0029] Step S105: Obtain the hardware condition parameters of the measurement system, including storage capacity and computing power, and evaluate the feasibility of the selected error correction strategy.

[0030] The hardware parameters of the measurement system, including storage capacity and computing power, are obtained and used as the basis for evaluating the feasibility of the error correction strategy. Based on the storage capacity, the amount of data the system can store is determined, and a suitable error correction strategy is selected accordingly. The computing power of the measurement system is analyzed to determine whether it meets the computational complexity requirements of the selected error correction strategy; if not, the strategy needs to be adjusted. A decision tree algorithm is used to automatically recommend the optimal error correction strategy, taking into account both storage capacity and computing power. The hardware configuration information of the measurement system is obtained and divided into different performance levels using a clustering algorithm, with a pre-defined applicable error correction strategy for each level. A mapping relationship between the measurement system hardware parameters and the error correction strategy is established, automatically adjusting the corresponding error correction strategy when hardware parameters change. An evaluation model of the matching degree between the measurement system hardware parameters and the error correction strategy is trained based on historical data using a support vector machine algorithm to guide the selection of the error correction strategy.

[0031] Specifically, to evaluate the feasibility of the error correction strategy, the hardware parameters of the measurement system need to be obtained first. By calling the system API, the system's storage capacity is found to be 500GB, and its computing power is 10 million data entries per second. Based on the storage capacity, the system can store approximately 500 million data entries. Based on this, a suitable error correction strategy can be selected, such as Hamming codes or Reed-Solomon codes. Simultaneously, by analyzing the computing power, it is found that the system's processing power of 10 million data entries per second fully meets the computational complexity requirements of Hamming codes. To automatically recommend the optimal error correction strategy, a decision tree algorithm can be used. Storage capacity and computing power are used as input features to the decision tree, and a decision tree model is generated by training on historical data. When new hardware parameters are input, the decision tree model can automatically recommend the optimal error correction strategy. Furthermore, clustering algorithms can be used to divide different hardware configurations into high, medium, and low performance levels. For each level, a suitable error correction strategy is preset, such as using Turbo codes for high-performance levels, LDPC codes for medium-performance levels, and BCH codes for low-performance levels. When hardware parameters change, the corresponding error correction strategy can be automatically adjusted according to a preset mapping relationship. Finally, an evaluation model can be trained using the Support Vector Machine (SVM) algorithm. The matching degree between hardware parameters and error correction strategies in historical data is used as training samples to obtain an evaluation model. When new hardware parameters are input, the evaluation model can predict the error correction strategy with the highest matching degree, thereby guiding the selection of the error correction strategy.

[0032] In step S106, if the hardware conditions meet the policy requirements, the corresponding error correction algorithm is loaded; if the hardware conditions are insufficient, the algorithm parameters are adjusted according to the priority to reduce the computational complexity.

[0033] Obtain performance metrics of the current hardware conditions, including memory capacity, processor frequency, and storage space, to form a hardware condition vector. Read the hardware requirements of each error correction algorithm from the policy library to form a policy requirement vector. Calculate the matching degree between the hardware condition vector and the policy requirement vector to obtain a fitness score for each error correction algorithm. Prioritize the error correction algorithms based on their fitness scores. Algorithms that meet the hardware conditions have higher priority, and those that do not have lower priority. If the hardware conditions meet the requirements of the highest priority error correction algorithm, load that algorithm directly and proceed to step 7; otherwise, proceed to step 6. For error correction algorithms that do not meet the hardware conditions, iterate through their parameters from highest to lowest priority, adjusting the parameters by reducing data precision and the number of iterations until the hardware conditions are met, obtaining a corrected set of algorithm parameters. Use the determined error correction algorithm and its parameter set to perform error correction processing on the received data and output the corrected data.

[0034] Specifically, the system first obtains the hardware parameters of the current device through the hardware detection module, such as 8GB of memory, 4GHz of processor frequency, and 256GB of storage space. These parameters are normalized to form a hardware condition vector [8,6,9]. Simultaneously, the system reads the hardware requirements of each error correction algorithm from the policy library. For example, algorithm A requires at least 4GB of memory and a processor frequency of at least 5GHz, while algorithm B requires at least 16GB of memory and a processor frequency of at least 2GHz. These requirements are quantified into policy requirement vectors, such as [[5,4], [0,8]]. Then, methods such as cosine similarity are used to calculate the matching degree between the hardware condition vector and each policy requirement vector, obtaining a fit score [9,3], with the corresponding algorithm priority ranking as [A,B]. Based on the calculation results, the current hardware is most suitable for algorithm A, and it can be directly loaded. If the device hardware conditions do not meet any algorithm requirements, the algorithm parameters are traversed according to priority, and adjustments are made by reducing data precision and the number of iterations. For example, reducing the data precision of Algorithm B from double-precision floating-point to single-precision can reduce memory usage by 50%, decrease the number of iterations from 10 to 5, and improve computation speed by 30%, thus meeting the requirements of current hardware conditions. Finally, the received data is corrected using the determined Algorithm A and its parameter set, outputting error-free data. Through this adaptive error correction algorithm selection mechanism, the optimal error correction algorithm can be dynamically selected and adjusted according to the actual hardware performance of the device, ensuring both error correction quality and system efficiency, and improving the stability and reliability of data transmission.

[0035] Step S107: During the data processing, the signal characteristics are monitored in real time. If the signal characteristics exceed the preset threshold, the operating condition type is reassessed.

[0036] Real-time signal data is acquired during data processing, and signal feature parameters are extracted. These extracted feature parameters are compared to preset thresholds to determine if they exceed the threshold range. If the feature parameters exceed the preset thresholds, a reassessment process for the operating condition type is triggered. Based on the signal feature change trend and historical operating condition types, a support vector machine algorithm is used to reassess the operating condition type. A decision tree algorithm, combined with the signal feature change characteristics and the assessed operating condition type, determines whether adjustments to the data processing flow are necessary. If adjustments are required, the parameters of the data processing algorithm are dynamically optimized based on the assessed operating condition type. The optimized data processing algorithm is applied to the real-time signal data to obtain processing results adapted to the current operating condition, thus completing the dynamic adjustment of the data processing flow.

[0037] Specifically, during data processing, the system acquires vibration signal data collected by sensors in real time and extracts characteristic parameters such as the root mean square (RMS) value and peak-to-peak value. These extracted parameters are compared with preset thresholds, for example, a RMS threshold of 5 m / s² and a peak-to-peak value threshold of 10 m / s². If a characteristic parameter exceeds the threshold range, a reassessment of the operating condition is triggered. Based on the signal characteristic change trend over the past 10 minutes and combined with historical operating condition data, the system uses a support vector machine algorithm to reassess the current operating condition, finding an 85% probability that the current operating condition is a "heavy load condition." Then, using a decision tree algorithm, the system comprehensively analyzes the signal characteristic change characteristics (such as a continuously rising RMS value) and the assessed operating condition type to determine the need for dynamic adjustments to the data processing flow. Based on the characteristics of a "heavy load condition," the system automatically adjusts the low-pass filter cutoff frequency in the data processing algorithm from 100Hz to 50Hz and the feature extraction time window from 1 second to 2 seconds, optimizing the parameter settings of the data processing algorithm. Finally, the optimized data processing algorithm was applied to real-time signal data to extract more accurate and stable feature parameters, obtain processing results adapted to the current heavy-load conditions, and complete the dynamic adjustment of the data processing flow, ensuring the reliability and effectiveness of data processing.

[0038] Step S108: Based on the re-evaluation results, dynamically adjust the error correction strategy to achieve an adaptive balance between error correction capability and real-time performance.

[0039] The process begins by acquiring the initial parameters and strategy of the error correction algorithm, processing the text data to be corrected, and obtaining the initial correction result. Based on preset evaluation metrics, the initial correction result is evaluated in real time, calculating metrics such as correction accuracy and speed to determine if it meets preset performance thresholds. If the real-time evaluation result does not meet the preset threshold, the parameters of the error correction algorithm are dynamically adjusted using machine learning algorithms, such as adjusting the correction granularity, correction rules, and feature weights, to generate a new error correction strategy. The dynamically adjusted error correction strategy is applied to the error correction algorithm to reprocess the text data, obtaining a new correction result. The new error correction result is evaluated in real time, calculating relevant performance metrics to determine if the adjusted error correction strategy effectively improves the error correction capability and real-time performance. If the new error correction strategy is effective, it is saved as the optimal error correction strategy and applied to subsequent error correction tasks; if ineffective, the process returns to step 3 to continue adjusting the error correction strategy. By iterating through steps 3-6, through continuous dynamic adjustment and evaluation optimization, the error correction algorithm achieves an adaptive balance between its error correction capability and real-time performance, meeting business requirements.

[0040] Specifically, when obtaining the initial parameters and strategies for the error correction algorithm, a statistical language model-based approach can be adopted. This involves training an N-gram model using a large-scale corpus and determining the initial error correction strategy by calculating the probability of candidate error-correcting words. For example, using 5 million Chinese sentences as training data, a 3-gram language model can be trained using the Kneser-Ney smoothing algorithm. The initial error correction granularity is set to the character level, and the error correction rules include homophones and similar-looking characters. Feature weights are set empirically. After applying the initial strategy to the error correction algorithm, 1000 texts to be corrected are processed. Through manual annotation, the initial error correction accuracy is calculated to be 82%, and the correction speed is 50ms / sentence. Based on a preset threshold, the initial results do not meet the 90% accuracy and 30ms / sentence speed requirements, necessitating optimization. By introducing a convolutional neural network (CNN) to automatically extract and combine error correction features, dynamically adjusting the error correction granularity to the word level, and dynamically updating the error correction rules and weights based on the error correction results, a new error correction strategy can be generated. After 10 rounds of iterative adjustments, the error correction accuracy was improved to 95%, and the error correction speed was optimized to 25ms / sentence, meeting the preset performance requirements. The optimized strategy was saved and applied to subsequent error correction tasks, achieving an adaptive balance between accuracy and speed in the error correction algorithm, effectively supporting business needs. In practical applications, the model can also be updated periodically using new error correction data through an incremental learning mechanism, enabling continuous optimization of the error correction strategy to adapt to constantly changing business scenarios.

[0041] Step S109: Output the error-corrected measurement data and record the strategy adjustment log for subsequent strategy optimization.

[0042] Acquire measurement data and perform error correction processing according to preset error correction rules to obtain corrected measurement data. Determine whether a strategy adjustment condition is triggered during the error correction process. If triggered, record a strategy adjustment log, including the adjusted strategy parameters and the reason for the adjustment. Output the corrected measurement data and save the strategy adjustment log as a reference for subsequent optimization of the error correction strategy. Train the optimized error correction strategy model using a machine learning algorithm on historical measurement data and corresponding error correction results. Apply the optimized error correction strategy model to new measurement data, perform error correction processing, and evaluate the error correction effect. Based on the error correction effect evaluation results, determine whether further optimization of the error correction strategy is needed. If so, return to step 4 to continue optimization. Solidify the finally optimized error correction strategy model and apply it to subsequent measurement data error correction processing. Continuously monitor the error correction effect and initiate a new round of strategy optimization if necessary.

[0043] Specifically, after acquiring measurement data, the system performs error correction processing according to preset error correction rules. For example, when outliers appear in the measurement data that exceed the normal range, the system automatically replaces them with the average of the two preceding and following normal values. Simultaneously, the system monitors the error correction process in real time. When more than five consecutive outliers occur, a strategy adjustment condition is triggered. The system automatically records the adjustment log, adjusts the outlier threshold to 80% of its original value, and records the reason for the adjustment as "too many consecutive outliers." The corrected measurement data is smoothed using a Kalman filter algorithm to remove high-frequency noise interference, resulting in a stable and reliable output. The system trains an optimized error correction strategy model using a gradient boosting decision tree algorithm on 10,000 sets of measurement data from the past month and their corresponding error correction results. This model is applied to 1,000 newly acquired sets of measurement data, and the error correction accuracy is improved from 95% to 98%. However, for some special operating conditions, the improvement in error correction effect is not significant. The system automatically triggers a strategy optimization process, further improving error correction performance through feature engineering and parameter tuning. The final optimized error correction strategy model is embedded into the system and applied to subsequent measurement data processing. Simultaneously, the system continuously monitors the error correction effect; if the error correction accuracy is below 95% for three consecutive days, a new round of strategy optimization is automatically initiated to ensure consistently high-quality measurement data.

[0044] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An adaptive error correction method in nuclear signal digital coincidence measurement, characterized in that, The method includes: Obtain the raw data stream from the digital coincidence measurement of the nuclear signal, extract the dynamic range and time-varying characteristic parameters of the signal, and establish a signal characteristic model; Based on the signal characteristic model, determine the type of operating condition the signal is in and determine the level of error correction requirement under the current operating condition; A multi-level error correction strategy library is pre-established, with each level of strategy corresponding to different error correction capabilities and computational complexity. The strategy library contains error correction algorithms for random errors and burst errors. Based on the current operating condition type and the level of error correction requirement, the optimal error correction strategy is matched from the strategy library to determine the type of error correction algorithm and parameter configuration to be used. During data processing, changes in signal characteristics are monitored in real time. If a signal characteristic is detected to exceed a preset threshold, the operating condition type is reassessed. Based on the reassessment results, the error correction strategy is dynamically adjusted to achieve an adaptive balance between error correction capability and real-time performance.

2. The method according to claim 1, characterized in that, The process of acquiring the raw data stream from the nuclear signal digital conformation measurement, extracting the signal's dynamic range and time-varying characteristic parameters, and establishing a signal characteristic model includes: Acquire the raw data stream of the nuclear signal, preprocess the data, and remove noise and interference information; Based on the time-domain and frequency-domain characteristics of the nuclear signal, the dynamic range and time-varying characteristic parameters of the signal are extracted; Feature selection is performed on the extracted characteristic parameters to identify key parameters that contribute to signal feature modeling; Machine learning algorithms, such as support vector machines or neural networks, are used to construct a feature model of nuclear signals; The performance of the established signal feature model is evaluated and optimized using cross-validation. The optimized nuclear signal feature model was applied to real data to assess its generalization ability and robustness. Based on the output of the nuclear signal characteristic model, the dynamic range and time-varying characteristics of the nuclear signal are determined, providing a basis for subsequent signal analysis.

3. The method according to claim 1, characterized in that, The step of determining the operating condition type of the signal based on the signal characteristic model and determining the error correction requirement level under the current operating condition includes: By extracting features from the signal, the feature vector of the signal is obtained, which serves as the basis for determining the type of operating condition of the signal. Based on the pre-established signal feature model, the extracted signal feature vector is input into the model, and the operating condition type of the signal is obtained through model reasoning. For different working conditions, different levels of error correction requirements are preset, and the corresponding error correction requirement level is obtained based on the working condition type determined by the model. If the matching degree between the feature vector of the signal and the feature model of any working condition type is lower than the preset threshold, it is determined to be an unknown working condition and the default error correction strategy is adopted. Based on the determined error correction requirement level, select the corresponding error correction strategy from the preset error correction strategy library and apply it to the error correction processing of the signal. During the training process of the signal feature model, the support vector machine algorithm is used to model the signal features under different working conditions, and the feature discrimination model for each working condition is obtained. When classifying error correction requirement levels, the signal quality requirements and fault tolerance capabilities under different operating conditions are comprehensively considered, and different error correction requirement levels and corresponding error correction strategies are set accordingly.

4. The method according to claim 1, characterized in that, The pre-established multi-level error correction strategy library, with each level corresponding to different error correction capabilities and computational complexity, includes error correction algorithms for random and burst errors, including: Based on different requirements for error correction capability and computational complexity, a corresponding multi-level error correction strategy library is established. Each level of the strategy library contains error correction algorithms for random errors and burst errors. Obtain the error type of the data to be corrected, determine whether it is a random error or a burst error, and select a suitable error correction algorithm from the corresponding error correction strategy library according to the error type. For different application scenarios and error correction requirements, select an error correction strategy from a multi-level error correction strategy library that meets the current error correction capability and computational complexity requirements; The data to be corrected is input into the selected error correction algorithm for processing. The error correction algorithm corrects the bit errors in the data to obtain the corrected data. The correctness of the corrected data is evaluated by comparing it with the original correct data and calculating the error correction rate of the error correction algorithm to verify the effectiveness of the selected error correction strategy. If the error correction rate does not meet expectations, a higher-level error correction strategy is selected from the multi-level error correction strategy library based on the trade-off between error correction rate and computational complexity, and the data is reprocessed for error correction. The above error correction process is iterated continuously until the corrected data meets the error correction requirements of the application scenario, and the final error correction result data is output.

5. The method according to claim 1, characterized in that, The step of matching the optimal error correction strategy from the strategy library based on the current operating condition type and error correction requirement level, and determining the type of error correction algorithm and parameter configuration to be used, includes: Obtain information on the current operating condition type and error correction requirement level as a basis for matching the optimal error correction strategy; In the preset strategy library, one or more error correction strategies are matched according to the current working condition type and the error correction requirement level. If multiple error correction strategies are matched, they are scored and ranked according to the preset evaluation rules, and the error correction strategy with the highest score is determined as the optimal error correction strategy. From the optimal error correction strategy, obtain information such as the error correction algorithm type and algorithm parameter configuration; Based on the type of error correction algorithm, the corresponding error correction algorithm model is called from the algorithm model library; Input the algorithm parameter configuration information into the error correction algorithm model to configure the parameters of the error correction algorithm model; Using a parameter-configured error correction algorithm model, the input data to be corrected is processed to obtain the corrected data output.

6. The method according to claim 1, characterized in that, After determining the type of error correction algorithm and parameter configuration, the hardware condition parameters of the measurement system are obtained, including storage capacity and computing power, and the feasibility of the selected error correction strategy is evaluated, including: Obtain the hardware parameters of the measurement system, including indicators such as storage capacity and computing power, and use them as the basis for evaluating the feasibility of the error correction strategy; Based on the storage capacity of the measurement system, determine the amount of data that the system can store, and select an appropriate error correction strategy accordingly. By analyzing the computational capabilities of the measurement system, it can be determined whether the system meets the computational complexity requirements of the selected error correction strategy. If not, the error correction strategy needs to be adjusted. The decision tree algorithm is used to automatically recommend the optimal error correction strategy by comprehensively considering the storage capacity and computing power of the measurement system. Obtain the hardware configuration information of the measurement system, divide it into different performance levels using a clustering algorithm, and pre-set an appropriate error correction strategy for each level; Establish a mapping relationship between measurement system hardware parameters and error correction strategies, and automatically adjust the corresponding error correction strategies when hardware parameters change; By using the support vector machine algorithm, an evaluation model for the matching degree between the hardware parameters of the measurement system and the error correction strategy is trained based on historical data, which can be used to guide the selection of the error correction strategy.

7. The method according to claim 6, characterized in that, After evaluating the feasibility of the selected error correction strategy, if the hardware conditions meet the strategy requirements, the corresponding error correction algorithm is loaded; if the hardware conditions are insufficient, the algorithm parameters are adjusted according to priority to reduce computational complexity, including: Obtain performance metrics of the current hardware conditions, including parameters such as memory capacity, processor frequency, and storage space, to form a hardware condition vector; Read the hardware requirements of each error correction algorithm from the policy library to form a policy requirement vector; The matching degree between the hardware condition vector and the policy requirement vector is calculated to obtain the fit score of each error correction algorithm. The error correction algorithms are prioritized based on their fit scores. Algorithms that meet the hardware requirements have higher priority, while those that do not have lower priority. If the hardware conditions meet the requirements of the highest priority error correction algorithm, then the algorithm is loaded directly. Otherwise, based on the current working condition type and error correction requirement level, the optimal error correction strategy is matched from the strategy library to determine the error correction algorithm type and parameter configuration to be used; For error correction algorithms that do not meet the hardware requirements, their algorithm parameters are traversed from high to low priority. The parameters are adjusted by reducing data precision and decreasing the number of iterations until the hardware requirements are met, thus obtaining the corrected algorithm parameter set. Using a defined error correction algorithm and its parameter set, the received data is processed to correct errors, and the corrected data is output.

8. The method according to claim 1, characterized in that, During data processing, real-time monitoring of signal characteristic changes is conducted. If a signal characteristic exceeds a preset threshold, the operating condition type is reassessed, including: Acquire real-time signal data during data processing and extract the characteristic parameters of the signal; The extracted signal feature parameters are compared with a preset threshold to determine whether they exceed the threshold range. If the signal characteristic parameters exceed the preset threshold, the re-evaluation process of the operating condition type will be triggered. Based on the trend of signal characteristic changes and historical operating condition types, the operating condition type is re-evaluated using the support vector machine algorithm; By using the decision tree algorithm, combined with the characteristics of signal feature changes and the type of operating condition being evaluated, it is determined whether the data processing flow needs to be adjusted. If the data processing workflow needs to be adjusted, the parameters of the data processing algorithm will be dynamically optimized according to the type of working condition being evaluated. The optimized data processing algorithm is applied to real-time signal data to obtain processing results adapted to the current working conditions, thus completing the dynamic adjustment of the data processing flow.

9. The method according to claim 1, characterized in that, The step of dynamically adjusting the error correction strategy based on the re-evaluation results to achieve an adaptive balance between error correction capability and real-time performance includes: Obtain the initial parameters and strategy of the error correction algorithm, perform error correction processing on the text data to be corrected, and obtain the initial error correction result; Based on the preset evaluation indicators, the initial error correction results are evaluated in real time, and indicators such as error correction accuracy and error correction speed are calculated to determine whether the preset performance threshold is met. If the real-time evaluation results do not meet the preset threshold, the parameters of the error correction algorithm are dynamically adjusted through machine learning algorithms based on the evaluation results, such as adjusting the error correction granularity, error correction rules and feature weights, to generate a new error correction strategy. The dynamically adjusted error correction strategy is applied to the error correction algorithm to reprocess the text data and obtain new error correction results. The new error correction results are evaluated in real time, relevant performance indicators are calculated, and it is determined whether the adjusted error correction strategy has effectively improved the error correction capability and real-time performance. If the new error correction strategy is effective, it is saved as the optimal error correction strategy and applied to subsequent error correction tasks. If ineffective, return to the pre-established multi-level error correction strategy library. Each level of strategy corresponds to different error correction capabilities and computational complexity. The strategy library contains error correction algorithms for random errors and burst errors, and the error correction strategy is adjusted again. Through continuous iteration and dynamic adjustments and evaluations, the error correction algorithm achieves an adaptive balance between its error correction capability and real-time performance, meeting business needs.

10. The method according to claim 1, characterized in that, After dynamically adjusting the error correction strategy, the corrected measurement data is output, and a strategy adjustment log is recorded for subsequent strategy optimization, including: Acquire measurement data, perform error correction processing on the measurement data according to preset error correction rules, and obtain the error-corrected measurement data; Determine whether the policy adjustment conditions are triggered during the error correction process. If so, record the policy adjustment log, including the adjusted policy parameters and the reason for the adjustment. Output the corrected measurement data and save the strategy adjustment log as a reference for subsequent optimization of the error correction strategy. The optimized error correction strategy model is obtained by training historical measurement data and corresponding error correction results using machine learning algorithms. The optimized error correction strategy model is applied to new measurement data to perform error correction and the error correction effect is evaluated. Based on the evaluation results of the error correction effect, determine whether further optimization of the error correction strategy is needed; The final optimized error correction strategy model is solidified and applied to subsequent measurement data error correction processing. The error correction effect is continuously monitored, and a new round of strategy optimization is initiated when necessary.

Images