A signal transmission data analysis system and method applied to an integrated circuit

Abstract

The application discloses a signal transmission data analysis system and method applied to an integrated circuit, relates to the technical field of transmission data analysis, collects circuit characteristics of a carrier carrying the integrated circuit, numbers the carrier, carries out interference analysis after classification according to fault types, removes invalid fault records caused by external interference, obtains reference fault records, extracts fault reference characteristic items corresponding to signal transmission faults based on the reference fault records, samples and detects the circuit characteristics of the carrier, combines the fault reference characteristics to evaluate the risk degree of various faults, switches other carriers to retransmit when it is determined that transmission risks exist and carries out comparison and verification, combines fault risks of the carrier to calculate transmission reliability, judges transmission true values through weighted voting, and accordingly labels normal samples and fault samples, and the application adopts a double interference removal mechanism combining time proximity preliminary screening and feature similarity fine correction, and distinguishes invalid alarms caused by external interference from real faults of the carrier itself.

Description

Technical Field

[0001] This invention relates to the field of transmission data analysis technology, specifically to a signal transmission data analysis system and method applied to integrated circuits. Background Technology

[0002] With the deep application of integrated circuits in key areas such as industrial control, communication hubs, and automotive electronics, the reliability of signal transmission is directly related to the operational safety and business continuity of the entire system. Data analysis and risk prevention and control of the signal transmission process have become core requirements in the field of integrated circuit operation and maintenance.

[0003] Existing integrated circuit signal transmission data analysis methods mostly involve collecting circuit characteristics of the carrier carrying the integrated circuit, extracting fault feature templates based on historical fault records, and performing real-time sampling and detection on the carrier in operation. Once a transmission risk is detected, the method switches to another carrier for retransmission, and the transmission correctness is verified by comparing the results of multiple transmissions. However, existing methods have significant technical shortcomings. Interference identification is based solely on the proximity of fault occurrence times, which cannot accurately distinguish between invalid alarms caused by external interference and genuine faults in the carrier itself. Either valid fault samples are mistakenly deleted, leading to distortion of fault characteristics, or interference samples are retained, reducing the accuracy of risk detection. Summary of the Invention

[0004] The purpose of this invention is to provide a signal transmission data analysis system and method for integrated circuits, so as to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a signal transmission data analysis method applied to integrated circuits, comprising the following steps: S1. Collect the circuit characteristics of the carriers carrying integrated circuits, number the carriers, and establish a carrier circuit database based on the circuit characteristics of the carriers in each factory area. S2. Retrieve historical alarm records from the plant area, classify them by fault type, perform interference analysis, eliminate invalid fault records caused by external interference, and obtain reference fault records. S3. Extract fault reference feature items corresponding to signal transmission faults based on reference fault records; S4. When the carrier receives a signal transmission task, sample and detect the circuit characteristics of the carrier, combine the fault reference characteristics to assess the risk level of various faults, and if it is determined that there is a transmission risk, switch to other carriers to retransmit and perform comparative verification. S5. For the transmission task that has been verified, calculate the transmission credibility based on the failure risk of the carrier, determine the transmission truth value through weighted voting, and label normal samples and faulty samples accordingly. S6. After each transmission cycle is completed, the operation log will be uploaded to the cloud platform to update the normal records, abnormal records, characteristic parameters and fault reference data. A credibility-weighted voting mechanism based on fault risk is adopted, and the feature parameters and fault reference data are dynamically adjusted through the cloud platform's periodic update mechanism to adapt to feature changes caused by carrier aging and environmental changes, upgrading passive post-event interference processing to active sample annotation and knowledge base iteration.

[0006] Furthermore, in step S1, after authorization, for any type of integrated circuit, data is collected from the carrier carrying the integrated circuit, each carrier is numbered, and the integrated circuit characteristics of each carrier are obtained. For any factory area, the integrated circuit characteristics of the carriers in the factory area are obtained, where the integrated circuit characteristics of the m-th carrier are {A(m_1), A(m_2), ..., A(m_n), ..., A(m_N)}, where m = 1, 2, ..., M, M is the number of carriers in the factory area, A(m_n) is the n-th integrated circuit characteristic of the m-th carrier, n = 1, 2, ..., N, N is the number of integrated circuit characteristics of the carriers. Substituting n = 1, 2, ..., N and m = 1, 2, ..., M one by one, all integrated circuit characteristics of all carriers in the factory area are obtained, and a carrier circuit database for any factory area is established based on all integrated circuit characteristics.

[0007] Further, in step S2, for any factory area, the historical log of the factory area is called to obtain the integrated circuit alarm record of the carrier. The integrated circuit alarm record includes the alarm carrier number, alarm time point and signal transmission fault type, wherein the number of signal transmission fault types is P. For the p-th signal transmission fault type, Q1 fault records of the p-th signal transmission fault type are obtained. For the Q1 fault records, the alarm time point of the Q1 fault records is called, and interference analysis conditions are set. The interference analysis conditions are that there is at least α fault records and another fault record exists, and the time interval between the two fault records is less than T1. The two fault records are classified into similar fault records at the corresponding time points, and the time points are recorded. Fault records that meet the interference analysis conditions are called, and the number of carriers corresponding to them is β. If β is greater than the preset threshold for the number of interference analysis carriers, the external interference judgment of the fault records is performed; otherwise, the number of fault records Q1 is used as the number of reference fault records Q. To determine external interference in fault records, for any recorded time point, the carriers corresponding to X similar fault records at that time point are called. The integrated circuit characteristics of the x-th carrier are {A(x_1), A(x_2), ..., A(x_n), ..., A(x_N)}, and the integrated circuit characteristics of the y-th carrier are {A(y_1), A(y_2), ..., A(y_n), ..., A(y_N)}. If there exists n such that ||A(x_n)-An| / An|>kn and ||A(y_n)-An| / An|>kn and ||A(x_n)-A(y_n)| / An|<kn, then the similar fault records corresponding to the x-th and y-th carriers are determined to be interference records. Where An is the average value of the nth integrated circuit feature of the carriers that have not experienced faults in the factory area, and kn is the feature difference threshold of the nth integrated circuit feature. This yields all interference records in the Q1 fault records. These interference records are then deleted from the Q1 fault records. After deletion, the number of reference fault records Q is obtained. Analysis is then conducted according to fault type. First, fault records are initially screened based on time similarity to quickly pinpoint the time range of potential interference, improving the efficiency of interference identification. Then, precise verification is performed by combining the similarity of the carrier's circuit features. This accurately distinguishes between invalid alarms caused by external interference and genuine faults of the carrier itself, avoiding both the distortion of fault features due to the accidental deletion of valid fault samples and the impact of residual interference samples on the accuracy of subsequent analysis. The reference fault records obtained through this double screening have higher purity, providing a reliable data foundation for subsequent fault reference feature extraction and ensuring the accuracy of risk detection from the source.

[0008] Further, in step S3, based on the reference fault records, the integrated circuit features of the carriers corresponding to the Q reference fault records at the time of the fault occurrence are called, where the q-th group of integrated circuit features is {A(p_q_1), A(p_q_2), ..., A(p_q_n), ..., A(p_q_N)}, and the n-th feature of the Q-th group of integrated circuit features is {A(p_1_n), A(p_2_n), ..., A(p_q_n), ..., A(p_Q_n)}. If ||A(p_q_n)-An| / An|>kn, then the n-th feature of the q-th group of integrated circuit features is determined to be a difference feature of the p-th signal transmission fault type; otherwise, the n-th feature of the q-th group of integrated circuit features is determined to be a non-difference feature of the p-th signal transmission fault type. Substituting each characteristic into q = 1, 2, ..., Q, we obtain the number of differential features for the p-th signal transmission fault type as Q(p_n). If Q(p_n) / Q is greater than K1, then we determine the fault reference feature item for the p-th signal transmission fault type based on the n-th feature, and thus obtain all fault reference feature items for the p-th signal transmission fault type. Similarly, for P signal transmission fault types, we obtain all fault reference feature items for any one of them. Based on the clean reference fault records filtered for interference, we extract the corresponding fault reference features according to the fault type. By statistically analyzing the frequency of occurrence of features in similar faults, we filter out core differential features, eliminating interference from occasional fluctuations and irrelevant features. The filtered fault reference features have a stronger correlation with the corresponding fault type and are more representative. Targeted feature sets can provide accurate judgment basis for subsequent real-time risk detection, improve the accuracy of fault risk identification, reduce the computational overhead of invalid features, reduce the computational burden of real-time detection, and balance accuracy and operational efficiency.

[0009] Furthermore, in step S4, when any carrier receives a signal transmission task, sampling is performed on any time point within the task's time period to obtain the integrated circuit characteristics of the carrier at that time point. For the p-th signal transmission fault type, all fault reference feature items for the p-th signal transmission fault type are obtained. The number of integrated circuit features corresponding to all fault reference feature items is D. Difference feature judgment is performed on the integrated circuit features corresponding to all fault reference feature items. The number of difference features among the integrated circuit features corresponding to all fault reference feature items is d. The p-th fault risk coefficient is Ep = d / D. If d / D is greater than Kp, it is determined that the carrier has a risk of the p-th signal transmission fault type; otherwise, the p-th fault risk coefficient is reset to 0, thereby obtaining P signal transmission faults of the carrier. The system uses a fault risk coefficient for each type of fault. If any fault risk coefficient is non-zero at any point in the time period, the signal transmission task is considered to have a transmission risk. An alternative carrier is then used to re-transmit the signal, and the signals from the two transmissions are compared for verification. If the signals are the same, the verification is considered successful; otherwise, another carrier is used to re-transmit the signal and the comparison continues until the signals are the same. Otherwise, the signal transmission task is considered to have no transmission risk. During signal transmission, the system uses sampling detection of carrier circuit characteristics, combined with pre-extracted fault reference features, to assess various fault risks in real time. This allows for risk assessment of all types of faults without interrupting normal transmission, balancing real-time detection with transmission efficiency. Determining risks by fault type accurately identifies specific fault categories, providing a clear basis for subsequent scheduling and processing. Once a transmission risk is determined, another carrier is immediately switched to re-transmit and the results are compared for verification. This timely avoidance of transmission risks caused by faulty carriers effectively ensures the correctness and reliability of signal transmission.

[0010] Furthermore, in step S5, for any task that completes signal transmission verification, the transmission results of all participating transmission carriers, the integrated circuit characteristics and fault risk coefficients at each sampling time point are obtained. The total number of carriers participating in this transmission verification is R, and the fault risk coefficient of the r-th carrier is E(r_p). The transmission reliability weight Wr of the r-th carrier is then calculated. ; All transmission results are weighted by credibility and voted on. The result with the highest number of votes and a cumulative weight exceeding a preset credibility threshold Wth is determined as the true value Str of this transmission. If the credibility threshold is not reached, additional carriers are called to participate in parallel transmission until the credibility threshold is met. Based on the true value, samples are labeled for all participating carriers: those with transmission results equal to Str are marked as normal samples, and those not equal to Str are marked as faulty samples, and the corresponding signal transmission fault type is matched. By combining the carrier fault risk to calculate the transmission credibility, a weighted voting method is used to determine the transmission truth value. This solves the verification failure problem caused by multiple carriers having the same error in traditional simple comparative verification, and significantly improves the accuracy of transmission correctness determination. The mechanism of supplementing parallel transmission carriers when the credibility threshold is not reached can flexibly adjust the verification redundancy, balancing reliability and operational efficiency. The automatic labeling of normal and faulty samples based on the true value can produce high-quality labeled data, providing reliable support for subsequent dynamic updates of the feature library and model iteration, and laying the core foundation for closed-loop optimization of the entire process.

[0011] Furthermore, in step S6, after each transmission cycle ends, the operation log is uploaded to the cloud platform, the normal records and abnormal records are updated, and the reference records corresponding to An and any fault type are updated.

[0012] A signal transmission data analysis system for integrated circuits, the system comprising: a circuit feature acquisition and database construction module, a fault record interference removal module, a fault reference feature extraction module, a transmission risk assessment and verification module, a transmission truth value determination and annotation module, and a cloud platform data periodic update module; The feature acquisition and on-chain module is used to collect information from both ends of the node, information about the data itself, and information about the link interference. The three types of information are preprocessed and normalized respectively, and the corresponding features are obtained and then synchronously stored in the blockchain distributed ledger. The effective feature filtering module is used to retrieve historical normal transmission samples archived on the chain, calculate the correlation between each feature and the transmission effect index, and filter out effective features that are significantly related to the transmission effect. The weight fitting modeling module is used to build a transmission effect analysis model based on historical samples, solve the independent influence weights of each effective feature on the transmission effect, reset the weights of unrelated features to zero, and generate a prediction model. The baseline interval calculation module is used to calculate the baseline transmission effect by substituting the corresponding features when receiving a data transmission task, and to calculate the interference intensity by combining the real-time link interference features, and to set a reasonable fluctuation range for the transmission effect. The transmission monitoring and tracing module is used to synchronously collect segmented data and store evidence during the data transmission process. After the transmission is completed, the measured transmission effect is compared with the reasonable fluctuation range to determine the transmission status. When there is an anomaly, the segmented data on the chain is retrieved to locate the source of the anomaly and synchronously fed back to all participating nodes. The cycle iteration optimization module is used to archive the transmission data and anomaly handling records of the current cycle onto the blockchain after each transmission cycle is completed, recalibrate the model parameters and judgment rules, and complete the system iteration optimization.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: On the one hand, this method employs a dual interference removal mechanism combining initial screening based on temporal proximity and fine-tuning based on feature similarity. This accurately distinguishes between invalid alarms caused by external interference and genuine faults of the carrier itself. It avoids both the erroneous deletion of valid fault samples leading to distortion of fault features and the prevention of residual interference samples that reduce the accuracy of risk detection. Furthermore, by selecting high-frequency difference features as fault reference features based on clean reference fault records, interference from occasional fluctuations and irrelevant features can be eliminated. The correlation and representativeness between fault features and corresponding fault types are stronger, thus improving the overall accuracy of fault risk detection from the source. On the one hand, this method calculates transmission reliability by combining carrier failure risk and uses a weighted voting method to determine the transmission truth value. The carrier with higher failure risk has a lower weight, which can effectively avoid the verification failure problem caused by multiple carriers having the same error and significantly improve the reliability of transmission correctness determination. At the same time, a flexible parallel carrier supplementation mechanism is adopted to dynamically add verification carriers when the reliability threshold is not reached. Under the premise of ensuring verification reliability, it avoids the waste of computing power and bandwidth caused by indiscriminate retransmission and balances transmission reliability and operating efficiency.

[0014] On the other hand, this method constructs a closed-loop self-evolving system covering the entire chain from risk detection and transmission verification to sample labeling and library updates. Based on transmission truth values, it automatically labels normal and faulty samples, and periodically updates normal feature benchmarks and fault reference feature data through a cloud platform. This dynamically adapts to feature shifts caused by carrier aging and changes in environmental temperature and humidity, avoiding the problem of continuous decline in detection accuracy over long-term operation. The entire process requires no manual intervention, enabling the system to self-iterate and self-optimize, and significantly improving long-term operational stability. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of a signal transmission data analysis system applied to integrated circuits according to the present invention; Figure 2 This is a flowchart of a signal transmission data analysis method for integrated circuits according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1 and Figure 2 The present invention provides a technical solution: a signal transmission data analysis method applied to integrated circuits, comprising the following steps: S1. Collect the circuit characteristics of the carriers carrying integrated circuits, number the carriers, and establish a carrier circuit database based on the circuit characteristics of the carriers in each factory area. S2. Retrieve historical alarm records from the plant area, classify them by fault type, perform interference analysis, eliminate invalid fault records caused by external interference, and obtain reference fault records. S3. Extract fault reference feature items corresponding to signal transmission faults based on reference fault records; S4. When the carrier receives a signal transmission task, sample and detect the circuit characteristics of the carrier, combine the fault reference characteristics to assess the risk level of various faults, and if it is determined that there is a transmission risk, switch to other carriers to retransmit and perform comparative verification. S5. For the transmission task that has been verified, calculate the transmission credibility based on the failure risk of the carrier, determine the transmission truth value through weighted voting, and label normal samples and faulty samples accordingly. S6. After each transmission cycle is completed, the operation log will be uploaded to the cloud platform to update the normal records, abnormal records, characteristic parameters and fault reference data. A credibility-weighted voting mechanism based on fault risk is adopted, and the feature parameters and fault reference data are dynamically adjusted through the cloud platform's periodic update mechanism to adapt to feature changes caused by carrier aging and environmental changes, upgrading passive post-event interference processing to active sample annotation and knowledge base iteration.

[0018] In step S1, after authorization, data is collected from the carriers carrying integrated circuits for any type of integrated circuit, each carrier is numbered, and the integrated circuit characteristics of each carrier are obtained. For any factory area, the integrated circuit characteristics of the carriers in the factory area are obtained, where the integrated circuit characteristics of the m-th carrier are {A(m_1), A(m_2), ..., A(m_n), ..., A(m_N)}, where m = 1, 2, ..., M, M is the number of carriers in the factory area, A(m_n) is the n-th integrated circuit characteristic of the m-th carrier, n = 1, 2, ..., N, N is the number of integrated circuit characteristics of the carriers. Substituting n = 1, 2, ..., N and m = 1, 2, ..., M into each carrier, all integrated circuit characteristics of all carriers in the factory area are obtained. Based on all integrated circuit characteristics, a carrier circuit database for any factory area is established.

[0019] In step S2, for any factory area, the historical log of the factory area is called to obtain the integrated circuit alarm record of the carrier. The integrated circuit alarm record includes the alarm carrier number, alarm time point and signal transmission fault type, wherein the number of signal transmission fault types is P. For the p-th signal transmission fault type, Q1 fault records of the p-th signal transmission fault type are obtained. For the Q1 fault records, the alarm time point of the Q1 fault records is called, and interference analysis conditions are set. The interference analysis conditions are that there is at least α fault records and another fault record exists, and the time interval between the two fault records is less than T1. The two fault records are classified into similar fault records at the corresponding time points, and the time points are recorded. Fault records that meet the interference analysis conditions are called, and the number of carriers corresponding to them is β. If β is greater than the preset threshold for the number of interference analysis carriers, the external interference judgment of the fault records is performed; otherwise, the number of fault records Q1 is used as the number of reference fault records Q. To determine external interference in fault records, for any recorded time point, the carriers corresponding to X similar fault records at that time point are called. The integrated circuit characteristics of the x-th carrier are {A(x_1), A(x_1), ..., A(x_n), ..., A(x_N)}, and the integrated circuit characteristics of the y-th carrier are {A(y_1), A(y_1), ..., A(y_n), ..., A(y_N)}. If there exists n such that ||A(x_n)-An| / An|>kn and || If A(y_n)-An| / An|>kn and ||A(x_n)-A(y_n)| / An|<kn, then the similar fault records corresponding to the x-th carrier and the y-th carrier are determined to be interference records, where An is the average value of the n-th integrated circuit feature of the carriers that have not experienced a fault in the factory area, and kn is the feature difference threshold of the n-th integrated circuit feature. Then, all interference records in the Q1 fault records are obtained, and the interference records are deleted from the Q1 fault records. After deletion, the number of reference fault records Q is obtained.

[0020] In step S3, based on the reference fault records, the integrated circuit features of the carriers corresponding to the Q reference fault records at the time of the fault occurrence are called, where the q-th group of integrated circuit features is {A(p_q_1), A(p_q_2), ..., A(p_q_n), ..., A(p_q_N)}, and the n-th feature of the Q groups of integrated circuit features is {A(p_1_n), A(p_2_n), ..., A(p_q_n), ..., A(p_Q_n)}. If ||A(p_q_n)-An| / An|>kn, then the n-th feature of the q-th group of integrated circuit features is determined to be the p-th feature. The difference features of the signal transmission fault type are determined; otherwise, the nth feature of the qth integrated circuit feature group is determined to be the indifferent feature of the pth signal transmission fault type. Substitute each feature into q=1,2,…,Q to obtain the number of difference features of the pth signal transmission fault type as Q(p_n). If Q(p_n) / Q is greater than K1, then the fault reference feature of the nth feature of the pth signal transmission fault type is determined, and then all fault reference feature items of the pth signal transmission fault type are obtained. Similarly, for P signal transmission fault types, all fault reference feature items of any one of them are obtained.

[0021] In step S4, when any carrier receives a signal transmission task, sampling is performed at any time point within the task's time period to obtain the integrated circuit characteristics of the carrier at that time point. For the p-th signal transmission fault type, all fault reference feature items for the p-th signal transmission fault type are obtained. The number of integrated circuit features corresponding to all fault reference feature items is D. Difference feature judgment is performed on the integrated circuit features corresponding to all fault reference feature items. The number of difference features among the integrated circuit features corresponding to all fault reference feature items is d. The p-th fault risk coefficient is Ep = d / D. If d / D is greater than Kp, it is determined that the carrier has a p-th signal transmission fault. If the risk of a fault is not zero at any point in time, the fault risk coefficient of the p-th fault is reset to 0, and the fault risk coefficients of the P signal transmission fault types of the carrier are obtained. If any fault risk coefficient is not zero at any point in time, it is determined that the signal transmission task has a transmission risk. Another carrier is called to re-perform the signal transmission task, and the signals after the two transmissions are compared and verified. If the signals after the transmissions are the same, the comparison and verification is determined to be successful. If the signals after the transmissions are different, another carrier is called to re-perform the signal transmission task and compare again until the signals after the transmissions are the same. Otherwise, it is determined that the signal transmission task has no transmission risk.

[0022] In step S5, for any task that completes signal transmission verification, the transmission results of all participating transmission carriers, the integrated circuit characteristics and fault risk coefficients at each sampling time point are obtained. The total number of carriers participating in this transmission verification is R, and the fault risk coefficient of the r-th carrier is E(r_p). The transmission reliability weight Wr of the r-th carrier is then calculated. ; All transmission results are weighted by credibility and voted on. The result with the highest number of votes and a cumulative weight exceeding the preset credibility threshold Wth is determined as the true value Str of this transmission. If the credibility threshold is not reached, additional carriers are called to participate in parallel transmission until the credibility threshold is met. Based on the true value, all participating carriers are labeled as samples: those with transmission results equal to Str are marked as normal samples, and those not equal to Str are marked as fault samples, and the corresponding signal transmission fault type is matched.

[0023] In step S6, after each transmission cycle is completed, the operation log is uploaded to the cloud platform, the normal records and abnormal records are updated, and the reference records corresponding to An and any fault type are updated.

[0024] A signal transmission data analysis system for integrated circuits includes: a circuit feature acquisition and library building module, a fault record interference removal module, a fault reference feature extraction module, a transmission risk assessment and verification module, a transmission truth value judgment and annotation module, and a cloud platform data periodic update module. The feature acquisition and on-chain module is used to collect information from both ends of the node, information about the data itself, and information about link interference. The three types of information are preprocessed and normalized respectively, and the corresponding features are obtained and then synchronously stored in the blockchain distributed ledger. The effective feature filtering module is used to retrieve historical normal transmission samples archived on the chain, calculate the correlation between each feature and the transmission effect index, and filter out effective features that are significantly related to the transmission effect. The weighted fitting modeling module is used to build a transmission effect analysis model based on historical samples, solve the independent influence weights of each effective feature on the transmission effect, reset the weights of unrelated features to zero, and generate a prediction model. The baseline interval calculation module is used to calculate the baseline transmission effect by substituting the corresponding features when a data transmission task is received, and to calculate the interference intensity by combining the real-time link interference features, and to set a reasonable fluctuation range for the transmission effect. The transmission monitoring and tracing module is used to synchronously collect and store segmented data during the data transmission process. After the transmission is completed, the measured transmission effect is compared with the reasonable fluctuation range to determine the transmission status. In case of an anomaly, the segmented data on the chain is retrieved to locate the source of the anomaly and synchronously fed back to all participating nodes. The periodic iteration optimization module is used to archive the transmitted data and anomaly handling records of each transmission cycle onto the blockchain after the end of each transmission cycle, recalibrate the model parameters and judgment rules, and complete the system iterative optimization.

[0025] Example 1: After authorization from the plant's operation and maintenance system, integrated circuit feature acquisition was performed on all 50 transmission carriers. Eight core features were collected for each carrier, including operating current, signal attenuation rate, clock offset, bit error rate, power supply ripple, temperature coefficient, impedance matching, and response delay. All features were normalized to dimensionless values ​​ranging from 0 to 1. Each of the 50 carriers was uniquely numbered, and steady-state feature data was collected for each carrier. The feature information of all carriers was integrated according to the plant area dimension, and the normalized mean values ​​of each feature for carriers that did not experience faults were calculated. For example, the normalized mean value of operating current was 0.42, and the normalized mean value of signal attenuation rate was 0.35. Finally, a carrier circuit database for the plant area was established. Historical alarm logs from the past 30 days were retrieved, revealing 12 records of "excessive bit error rate" faults. With a time interval threshold of 5 minutes, analysis of the alarm timestamps of 12 faults revealed that 6 faults occurred within the same 10-minute period, involving 4 different carriers. The number of carriers exceeded the preset threshold of 3 for interference analysis, thus triggering external interference detection. The operating current characteristics of the 4 carriers during this period were retrieved: the normalized operating current values ​​at the fault time for two carriers were 0.68 and 0.69, respectively. The deviations from the normal average were calculated as (0.68-0.42) / 0.42≈0.619 and (0.69-0.42) / 0.42≈0.643, both exceeding the preset characteristic difference threshold of 0.2. The relative difference between the two characteristics was (0.69-0.68) / 0.42≈0.024, less than 0.2, therefore these two faults were determined to be external interference records. Two interference records were removed from the 12 faults, resulting in 10 reference fault records.

[0026] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A signal transmission data analysis method applied to integrated circuits, characterized in that: The method includes the following steps: S1. Collect the circuit characteristics of the carriers carrying integrated circuits, number the carriers, and establish a carrier circuit database based on the circuit characteristics of the carriers in each factory area. S2. Retrieve historical alarm records from the plant area, classify them by fault type, perform interference analysis, eliminate invalid fault records caused by external interference, and obtain reference fault records. S3. Extract fault reference feature items corresponding to signal transmission faults based on reference fault records; S4. When the carrier receives a signal transmission task, sample and detect the circuit characteristics of the carrier, combine the fault reference characteristics to assess the risk level of various faults, and if it is determined that there is a transmission risk, switch to other carriers to retransmit and perform comparative verification. S5. For the transmission task that has been verified, calculate the transmission credibility based on the failure risk of the carrier, determine the transmission truth value through weighted voting, and label normal samples and faulty samples accordingly. S6. After each transmission cycle is completed, the operation log will be uploaded to the cloud platform, and the normal records, abnormal records, characteristic parameters and fault reference data will be updated.

2. The signal transmission data analysis method for integrated circuits according to claim 1, characterized in that: A credibility-weighted voting mechanism based on fault risk is adopted, and the feature parameters and fault reference data are dynamically adjusted through the cloud platform's periodic update mechanism to adapt to feature changes caused by carrier aging and environmental changes, upgrading passive post-event interference processing to active sample annotation and knowledge base iteration.

3. The signal transmission data analysis method applied to integrated circuits according to claim 2, characterized in that: In step S1, after authorization, data is collected from the carriers carrying integrated circuits for any type of integrated circuit, each carrier is numbered, and the integrated circuit characteristics of each carrier are obtained. For any factory area, the integrated circuit characteristics of the carriers in the factory area are obtained, where the integrated circuit characteristics of the m-th carrier are {A(m_1), A(m_2), ..., A(m_n), ..., A(m_N)}, where m = 1, 2, ..., M, M is the number of carriers in the factory area, A(m_n) is the n-th integrated circuit characteristic of the m-th carrier, n = 1, 2, ..., N, N is the number of integrated circuit characteristics of the carriers. Substituting n = 1, 2, ..., N and m = 1, 2, ..., M into each carrier, all integrated circuit characteristics of all carriers in the factory area are obtained, and a carrier circuit database for any factory area is established based on all integrated circuit characteristics.

4. The signal transmission data analysis method for integrated circuits according to claim 3, characterized in that: In step S2, for any factory area, the historical log of the factory area is called to obtain the integrated circuit alarm record of the carrier. The integrated circuit alarm record includes the alarm carrier number, alarm time point and signal transmission fault type, wherein the number of signal transmission fault types is P. For the p-th signal transmission fault type, Q1 fault records of the p-th signal transmission fault type are obtained. For the Q1 fault records, the alarm time point of the Q1 fault records is called, and interference analysis conditions are set. The interference analysis conditions are that there is at least α fault records and another fault record exists, and the time interval between the two fault records is less than T1. The two fault records are classified into similar fault records at the corresponding time points, and the time points are recorded. Fault records that meet the interference analysis conditions are called, and the number of carriers corresponding to them is β. If β is greater than the preset threshold for the number of interference analysis carriers, the external interference judgment of the fault records is performed; otherwise, the number of fault records Q1 is used as the number of reference fault records Q. To determine external interference in fault records, for any recorded time point, the carriers corresponding to X similar fault records at that time point are called. The integrated circuit characteristics of the x-th carrier are {A(x_1), A(x_1), ..., A(x_n), ..., A(x_N)}, and the integrated circuit characteristics of the y-th carrier are {A(y_1), A(y_1), ..., A(y_n), ..., A(y_N)}. If there exists n such that ||A(x_n)-An| / An|>kn and || If A(y_n)-An| / An|>kn and ||A(x_n)-A(y_n)| / An|<kn, then the similar fault records corresponding to the x-th carrier and the y-th carrier are determined to be interference records, where An is the average value of the n-th integrated circuit feature of the carriers that have not experienced a fault in the factory area, and kn is the feature difference threshold of the n-th integrated circuit feature. Then, all interference records in the Q1 fault records are obtained, and the interference records are deleted from the Q1 fault records. After deletion, the number of reference fault records Q is obtained.

5. The signal transmission data analysis method for integrated circuits according to claim 4, characterized in that: In step S3, based on the reference fault records, the integrated circuit features of the carriers corresponding to the Q reference fault records at the time of the fault occurrence are called, where the q-th group of integrated circuit features is {A(p_q_1), A(p_q_2), ..., A(p_q_n), ..., A(p_q_N)}, and the n-th feature of the Q groups of integrated circuit features is {A(p_1_n), A(p_2_n), ..., A(p_q_n), ..., A(p_Q_n)}. If ||A(p_q_n)-An| / An|>kn, then the n-th feature of the q-th group of integrated circuit features is determined to be the p-th feature. The difference features of the signal transmission fault type are determined; otherwise, the nth feature of the qth integrated circuit feature group is determined to be the indifferent feature of the pth signal transmission fault type. Substitute each feature into q=1,2,…,Q to obtain the number of difference features of the pth signal transmission fault type as Q(p_n). If Q(p_n) / Q is greater than K1, then the fault reference feature of the nth feature of the pth signal transmission fault type is determined, and then all fault reference feature items of the pth signal transmission fault type are obtained. Similarly, for P signal transmission fault types, all fault reference feature items of any one of them are obtained.

6. The signal transmission data analysis method applied to integrated circuits according to claim 5, characterized in that: In step S4, when any carrier receives a signal transmission task, sampling is performed at any time point within the task's time period to obtain the integrated circuit characteristics of the carrier at that time point. For the p-th signal transmission fault type, all fault reference feature items for the p-th signal transmission fault type are obtained. The number of integrated circuit features corresponding to all fault reference feature items is D. Difference feature judgment is performed on the integrated circuit features corresponding to all fault reference feature items. The number of difference features among the integrated circuit features corresponding to all fault reference feature items is d. The p-th fault risk coefficient is Ep = d / D. If d / D is greater than Kp, it is determined that the carrier has a p-th signal transmission fault. If the risk of a fault is not zero at any point in time, the fault risk coefficient of the p-th fault is reset to 0, and the fault risk coefficients of the P signal transmission fault types of the carrier are obtained. If any fault risk coefficient is not zero at any point in time, it is determined that the signal transmission task has a transmission risk. Another carrier is called to re-perform the signal transmission task, and the signals after the two transmissions are compared and verified. If the signals after the transmissions are the same, the comparison and verification is determined to be successful. If the signals after the transmissions are different, another carrier is called to re-perform the signal transmission task and compare again until the signals after the transmissions are the same. Otherwise, it is determined that the signal transmission task has no transmission risk.

7. The signal transmission data analysis method for integrated circuits according to claim 6, characterized in that: In step S5, for any task that completes signal transmission verification, the transmission results of all participating transmission carriers, the integrated circuit characteristics and fault risk coefficients at each sampling time point are obtained. The total number of carriers participating in this transmission verification is R, and the fault risk coefficient of the r-th carrier is E(r_p). The transmission reliability weight Wr of the r-th carrier is then calculated. ； All transmission results are weighted by credibility and voted on. The result with the highest number of votes and a cumulative weight exceeding the preset credibility threshold Wth is determined as the true value Str of this transmission. If the credibility threshold is not reached, additional carriers are called to participate in parallel transmission until the credibility threshold is met. Based on the true value, all participating carriers are labeled as samples: those with transmission results equal to Str are marked as normal samples, and those not equal to Str are marked as fault samples, and the corresponding signal transmission fault type is matched.

8. The signal transmission data analysis method for integrated circuits according to claim 7, characterized in that: In step S6, after each transmission cycle is completed, the operation log is uploaded to the cloud platform, the normal records and abnormal records are updated, and the reference records corresponding to An and any fault type are updated.

9. A signal transmission data analysis system for integrated circuits, wherein the system is applied to the signal transmission data analysis method for integrated circuits as described in any one of claims 1-8, characterized in that: The system includes: a circuit feature acquisition and database construction module, a fault record interference removal module, a fault reference feature extraction module, a transmission risk assessment and verification module, a transmission truth value judgment and annotation module, and a cloud platform data periodic update module. The circuit feature acquisition and database building module is used to collect the circuit features of carriers carrying integrated circuits, number the carriers, and establish a carrier circuit database based on the circuit features of carriers in each factory area. The fault record interference removal module is used to retrieve historical alarm records of the plant area, classify them according to fault type, perform interference analysis, remove invalid fault records caused by external interference, and obtain reference fault records. The fault reference feature extraction module is used to extract fault reference feature items corresponding to signal transmission faults based on reference fault records; The transmission risk assessment and verification module is used to sample and detect the circuit characteristics of the carrier when the carrier receives a signal transmission task, assess the risk level of various faults in combination with fault reference characteristics, and switch to other carriers to retransmit and perform comparative verification when it is determined that there is a transmission risk. The transmission truth value determination and labeling module is used to complete the verification transmission task. It calculates the transmission credibility based on the carrier's failure risk, determines the transmission truth value through weighted voting, and labels normal samples and faulty samples accordingly. The cloud platform data periodic update module is used to upload the operation log to the cloud platform after each transmission work cycle, and update the normal records, abnormal records, characteristic parameters and fault reference data.

Images