AI dynamic monitoring guard system based on edge calculation and implementation method

Abstract

The invention discloses an AI dynamic monitoring guard system based on edge computing and an implementation method, relates to the technical field of highway toll collection equipment monitoring, and solves the technical problem that the running states of hardware equipment and software service cannot be monitored at the same time in the prior art. The method comprises the following steps: acquiring safety data of a highway toll collection system and environment data corresponding to a collection moment; obtaining a correlation coefficient of the operation data and the log data through calculation, and processing the security data based on the correlation coefficient to obtain equipment target data and software target data; calculating an equipment fault evaluation coefficient and a software fault evaluation coefficient based on the equipment target data and the software target data; calculating to obtain a safety evaluation coefficient based on the equipment fault evaluation coefficient and the software fault evaluation coefficient, and correcting the safety evaluation coefficient based on the environmental data to obtain a target safety evaluation coefficient; and judging whether to carry out early warning based on the target safety evaluation coefficient. The technical problem is solved.

Description

Technical Field

[0001] The present invention belongs to the field of monitoring of highway toll collection equipment, and specifically relates to an AI dynamic monitoring and guarding system based on edge computing and an implementation method thereof. Background Art

[0002] The monitoring and guarding of highway toll collection equipment and software systems are important components of the operation and management of toll stations; traditional technologies use networks, sensors, and edge data processing technologies to achieve real-time monitoring of equipment and software, mainly by receiving data sent by the equipment to determine whether the equipment is online; when a fault occurs in the equipment, it is necessary for mechanical and electrical management personnel to conduct regular inspections and solve problems on-site.

[0003] Existing monitoring systems mainly achieve independent monitoring of hardware devices or software services by receiving data sent by the devices, with relatively single functions, only being able to display the online or offline status of the devices, and unable to simultaneously take into account the operating conditions of both hardware and software; such systems rely on mechanical and electrical management personnel to conduct regular equipment inspections, and manual on-site problem-solving is required when a fault occurs, resulting in low response efficiency. In addition, due to the lack of in-depth data analysis and intelligent processing capabilities, existing technologies are unable to predict potential faults and can only passively discover and handle problems after they occur, which not only affects the continuity of operation but also increases maintenance costs and management complexity.

[0004] Therefore, the present invention proposes an AI dynamic monitoring and guarding system based on edge computing and an implementation method thereof to solve the above problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; for this purpose, the present invention proposes an AI dynamic monitoring and guarding system based on edge computing and an implementation method thereof, which are used to solve the technical problems that the prior art can only independently monitor hardware devices or software services and cannot simultaneously take into account the operating states of both; and that the monitoring system only displays the online or offline status of the devices, requires manual on-site handling of faults, has low response efficiency; and is unable to predict potential faults through data analysis, resulting in passive discovery of faults and affecting the continuity of operation.

[0006] To achieve the above object, the first aspect of the present invention provides an AI dynamic monitoring and guarding system based on edge computing, including: a data acquisition module, a data analysis module, and an early warning module;

[0007] The data acquisition module: is used to obtain the security data of the highway toll collection system and the environmental data at the corresponding acquisition moment; wherein, the security data includes: the operation data of the equipment and the log data of the software;

[0008] Data analysis module: Calculate the correlation coefficient of the operation data and the log data, and process the security data based on the correlation coefficient to obtain the device target data and the software target data; and,

[0009] Calculate the device failure evaluation coefficient and the software failure evaluation coefficient based on the device target data and the software target data; calculate the security evaluation coefficient based on the device failure evaluation coefficient and the software failure evaluation coefficient, and correct the security evaluation coefficient based on the environmental data to obtain the target security evaluation coefficient;

[0010] Early warning module: Judge whether to give an early warning based on the target security evaluation coefficient.

[0011] Preferably, the obtaining of the security data of the highway toll collection system and the environmental data at the corresponding collection time includes:

[0012] Real-time collect the operation data of the highway toll collection equipment through data sensors; among them, the operation data includes: the temperature, current, voltage, output power, vibration amplitude, noise level, operation time, etc. of the equipment;

[0013] Collect the log data of the highway toll collection software system through data collection equipment; among them, the data collection equipment includes: Fluentd tool or Prometheus tool; the log data includes: API response time, CPU usage rate, memory usage rate, network throughput, log error rate, network latency, etc.;

[0014] Real-time collect the environmental data through data sensors; among them, the environmental data includes: environmental temperature and environmental humidity.

[0015] It should be noted that the operation data refers to the various index data obtained in real time through data collection equipment, reflecting the physical state and operation performance of the highway toll collection equipment. The operation data mainly comes from sensors, monitoring systems or equipment interfaces, and can help judge the health status and potential failure risks of the equipment; the log data refers to the structured or unstructured record information generated from the software system, used to describe the events, operations and performance manifestations occurring during the operation of the system. The log data is mainly used to analyze the operation status, performance bottlenecks and abnormal behaviors of the software system, and provide a basis for optimizing the system performance and troubleshooting problems.

[0016] Preferably, the calculating the correlation coefficient of the operation data and the log data includes: calculating the correlation coefficient of the operation data and calculating the correlation coefficient of the log data;

[0017] The method for calculating the correlation coefficient of the operation data is:

[0018] A1: Obtain several sets of operation data, log data, the device status corresponding to each data in the operation data, and the software status corresponding to each data in the log data from the historical database of the highway toll collection system; wherein, the device status includes: device failure and normal device; the software status includes: software failure and normal software.

[0019] A2: Mark several sets of operation data as Y i,g ; wherein, Y i,g refers to the i-th data in the g-th set of operation data, g = {1, 2, 3, …, M}, M is the total number of sets, i = {1, 2, 3, …, N}, and N is the total number of data included in the operation data.

[0020] A3: Calculate the probability that each data in the operation data causes device failure through the formula P i = G i / g; wherein, P i refers to the probability that the i-th data in the operation data causes device failure, i is the label of each data in the operation data, and G i refers to the total number of times that the i-th data in the operation data causes device failure.

[0021] A4: Obtain the correlation coefficient between each data in the operation data and device failure through the formula XG i = (e ^ (P i ) - e ^ (-P i )) / (e ^ (P i ) + e ^ (-P i )); wherein, XG i refers to the correlation coefficient between the i-th data in the operation data and device failure.

[0022] It should be noted that the larger the value of the correlation coefficient, the stronger the correlation with the failure event; the historical database of the highway toll collection system refers to the database that stores all historical operation data, log records, and related status information related to the highway toll collection system.

[0023] Preferably, calculating the correlation coefficient of the log data includes:

[0024] B1: Mark several sets of log data as R j,g ; wherein, R j,g refers to the j-th data in the g-th set of log data, j = {1, 2, 3, …, L}, and L is the total number of data included in the log data.

[0025] B2: Calculate the probability that each data in the log data causes software failure through the formula P j = G j / g; wherein, P jRefers to the probability that the j-th data in the log data causes a software failure, where j is the label of each data in the log data, G j Refers to the total number of times that the j-th data in the log data causes a software failure;

[0026] B3: Through the formula XG j =(e^(P i )) - e^(-P i )) / (e^(P i )) + e^(-P i )) to obtain the correlation coefficient between each data in the log data and the software failure; where XG j Refers to the correlation coefficient between the j-th data in the log data and the software failure.

[0027] Preferably, the processing of the security data based on the correlation coefficient includes:

[0028] Judging whether the correlation coefficient between each data in the operation data and the equipment failure and the correlation coefficient between each data in the log data and the software failure are greater than a preset correlation coefficient threshold; if yes, retain the corresponding data; if not, remove the corresponding data;

[0029] Mark the retained operation data as equipment target data and mark the retained log data as software target data.

[0030] It should be noted that the preset correlation coefficient threshold is set by experts in the field according to actual security requirements. The higher the actual security requirements, the larger the preset correlation coefficient threshold.

[0031] Preferably, the calculation of the equipment failure evaluation coefficient and the software failure evaluation coefficient based on the equipment target data and the software target data includes: calculating the equipment failure evaluation coefficient based on the equipment target data and calculating the software failure evaluation coefficient based on the software target data;

[0032] The calculation of the equipment failure evaluation coefficient based on the equipment target data includes:

[0033] Extracting the equipment target data and the corresponding correlation coefficient; marking the equipment target data as SM e and marking the correlation coefficient corresponding to the equipment target data as XG e ; where SM e Refers to the e-th data in the equipment target data, and XG e Refers to the correlation coefficient corresponding to the e-th data in the equipment target data, where e = {1, 2, 3,..., Q} and Q is the total number of data included in the equipment target data;

[0034] Through the formula x = ∑(1 / XG e)Calculate the device weight factor; where x is the device weight factor, ∑ is the summation symbol, and the summation range is e;

[0035] Through the formula α e = 1 / (XG e ×x) to calculate the weight coefficient corresponding to each data in the device target data; where α e refers to the weight coefficient corresponding to the e-th data in the device target data;

[0036] Through the formula P1 = ∑(α e ×e^((SM e -ZS e ) / ZS e )) to calculate the device fault evaluation coefficient; where P1 is the device fault evaluation coefficient, ∑ is the summation symbol, the summation range is e, and ZS e refers to the optimal data corresponding to the e-th data in the device target data.

[0037] It should be noted that the optimal data corresponding to the e-th data in the device target data refers to the best value or reference value of this data under the ideal operating state or normal operating state of the device.

[0038] Preferably, calculating the software fault evaluation coefficient based on the software target data includes:

[0039] Extract the software target data and the corresponding correlation coefficients; mark the software target data as RM k , and mark the correlation coefficient corresponding to the software target data as RX k ; where RM k refers to the k-th data in the software target data, and RX k refers to the correlation coefficient corresponding to the k-th data in the software target data, k = {1, 2, 3,..., V}, and V is the total number of data included in the software target data;

[0040] Through the formula y = ∑(1 / RX k ) to calculate the software weight factor; where y is the software weight factor, ∑ is the summation symbol, and the summation range is k;

[0041] Through the formula β k = 1 / (RX k ×y) to calculate the weight coefficient corresponding to each data in the software target data; where β k refers to the weight coefficient corresponding to the k-th data in the software target data;

[0042] Through the formula P2 = ∑(β k ×e^((RM k -ZR k) / ZR k )) Calculate the software failure evaluation coefficient; where P2 refers to the software failure evaluation coefficient, ∑ is the summation symbol, and the summation range is k, and ZR k refers to the optimal data corresponding to the k-th data in the software target data.

[0043] It should be noted that the optimal data corresponding to the k-th data in the software target data refers to the best value or reference value of this data when the software is in an ideal operating state or a normal operating state.

[0044] Preferably, the correction of the security evaluation coefficient based on environmental data includes:

[0045] Extract the equipment failure evaluation coefficient P1 and the software failure evaluation coefficient P2;

[0046] Calculate the security evaluation coefficient through the formula P = θ1×P1 + θ2×P2; where P is the security evaluation coefficient, and θ1 and θ2 are the initial weight coefficients;

[0047] Obtain the highest environmental temperature, the lowest environmental temperature, the highest environmental humidity, and the lowest environmental humidity of the day based on the weather prediction platform, and mark them as Tmax, Tmin, Hmax, and Hmin respectively;

[0048] Calculate the environmental temperature index through the formula Tn = (T - Tmin) / (Tmax - Tmin); where Tn is the environmental temperature index and T is the environmental temperature;

[0049] Calculate the environmental humidity index through the formula Hn = (H - Hmin) / (Hmax - Hmin); where Hn is the environmental humidity index and H is the environmental humidity;

[0050] Correct the initial weight coefficient θ1 based on the environmental temperature index, specifically:

[0051] Calculate the corrected weight coefficient of θ1 through the formula θ1’ = θ1 + f1×Tn - f2×Hn; where θ1’ is the corrected weight coefficient of θ1, f1 is the temperature influence factor, and f2 is the humidity influence factor;

[0052] Correct the initial weight coefficient θ2 based on the environmental humidity index, specifically:

[0053] Calculate the corrected weight coefficient of θ2 through the formula θ2’ = θ2 - f1×Tn + f2×Hn; where θ2’ is the corrected weight coefficient of θ2;

[0054] Obtain the target weight coefficients by normalizing the corrected weight coefficients θ1’ and θ2’, and the normalization formula is specifically:

[0055] ∈1 = θ1’ / (θ1’ + θ2’) and ∈2 = θ2’ / (θ1’ + θ2’); where ∈1 and ∈2 are target weight coefficients;

[0056] Update the initial weight coefficients θ1 and θ2 to ∈1 and ∈2 respectively;

[0057] Calculate the target safety assessment coefficient through the formula P’ = ∈1 × P1 + ∈2 × P2; where, P’ is the target safety assessment coefficient.

[0058] It should be noted that the initial weight coefficients are set by experts in this field according to actual experience. In historical maintenance data, if the impact degree of equipment failure is greater, then θ1 > θ2;

[0059] The temperature impact factor reflects the sensitivity of temperature to equipment failure. If historical data shows that for every 10°C increase in temperature, the equipment failure rate increases by 20%, then set f1 = 0.2; the humidity impact factor reflects the sensitivity of humidity to software anomalies; both the temperature impact factor and the humidity impact factor can be obtained from the historical database of the highway toll system and the corresponding historical environmental data analysis.

[0060] Under high - temperature environments, the equipment failure risk rises significantly. It is necessary to ensure that the system gives priority to the health status of hardware by increasing the equipment evaluation weight; high humidity can cause circuit board leakage or short - circuit, affecting the stability of network equipment. At the same time, high - frequency signals attenuate more severely in humid air, leading to an increase in network latency or packet loss rate. By increasing the software evaluation weight, the system pays more attention to network metrics (such as latency, packet loss rate) in log data and captures software anomalies in a timely manner.

[0061] Preferably, determining whether to give an early warning based on the target safety assessment coefficient includes:

[0062] Judge whether the target safety assessment coefficient is greater than the preset target safety assessment coefficient threshold; if yes, generate a warning message and send it to the client; if no, continue to detect and judge.

[0063] It should be noted that the preset target safety assessment coefficient threshold is set by experts in this field according to actual safety requirements. The higher the actual safety requirements, the larger the preset target safety assessment coefficient threshold.

[0064] The second aspect of the present invention provides an AI dynamic monitoring and guarding method based on edge computing, including:

[0065] Step 1: Obtain the safety data of the highway toll system and the environmental data at the corresponding acquisition moment;

[0066] Step 2: Calculate the correlation coefficient of the operation data and the log data, and process the security data based on the correlation coefficient to obtain the device target data and the software target data;

[0067] Step 3: Calculate the device failure evaluation coefficient and the software failure evaluation coefficient based on the device target data and the software target data;

[0068] Step 4: Calculate the security evaluation coefficient based on the device failure evaluation coefficient and the software failure evaluation coefficient, and correct the security evaluation coefficient based on the environmental data to obtain the target security evaluation coefficient;

[0069] Step 5: Determine whether to give an early warning based on the target security evaluation coefficient.

[0070] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0071] 1. The present invention obtains the security data of the highway toll collection system and the environmental data at the corresponding collection moment; calculates the correlation coefficient of the operation data and the log data, and processes the security data based on the correlation coefficient to obtain the device target data and the software target data; calculates the device failure evaluation coefficient and the software failure evaluation coefficient based on the device target data and the software target data; calculates the security evaluation coefficient based on the device failure evaluation coefficient and the software failure evaluation coefficient, and corrects the security evaluation coefficient based on the environmental data to obtain the target security evaluation coefficient; determines whether to give an early warning based on the target security evaluation coefficient, solving the technical problems that the prior art can only independently monitor hardware devices or software services, cannot take into account the operating states of both at the same time; and the monitoring system only displays the online or offline state of the device, requires manual on-site handling of faults, and has low response efficiency; cannot predict potential faults through data analysis, resulting in passive discovery of faults and affecting the operation continuity.

[0072] 2. The present invention can quantitatively analyze the correlation between device operation data and device failures, and between software log data and software failures, so as to screen out key data with high contribution to fault prediction; by using the rich data resources in the historical database, combining probability statistics methods and mathematical formulas, calculate the correlation coefficient between each operation data and log data and the fault event, and the larger the correlation coefficient value, the stronger its correlation with the fault; on this basis, by setting a preset correlation coefficient threshold, further screen out the device target data and software target data highly related to the fault, effectively reducing the interference of redundant data, improving the accuracy of subsequent evaluations. This method can not only accurately locate potential fault factors, but also provide a scientific data basis for the system, significantly improving the fault prediction ability and the intelligent level of the system, while reducing unnecessary calculation costs and complexities, laying a solid foundation for the efficient operation and maintenance and safety management of the highway toll collection system.

[0073] 3. The present invention calculates the device failure evaluation coefficient and the software failure evaluation coefficient based on the device target data and the software target data respectively, which can comprehensively evaluate the operating status of the hardware device and the software system, and dynamically correct the overall security evaluation coefficient in combination with the environmental data, significantly improving the accuracy and intelligent level of fault prediction; by introducing weight factors and exponential functions into the device target data and the software target data, the importance of different data and the degree of deviation from the ideal state are fully considered, making the evaluation results more scientific and reasonable; by dynamically adjusting the initial weight coefficient through the environmental temperature and humidity index, the actual impact of the external environment on the operating status of the device and the software is reflected, ensuring that the system can still accurately warn of potential risks under extreme weather conditions; by setting a preset target security evaluation coefficient threshold to achieve the automatic warning function, not only the need for manual intervention is reduced, but also the response efficiency is greatly improved, providing a reliable guarantee for the stable operation of the highway toll collection system, effectively integrating multi-source data, taking into account the comprehensive evaluation requirements of hardware and software, greatly enhancing the adaptability and robustness of the system, and laying a solid foundation for realizing intelligent operation and maintenance management. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0075] Figure 1 It is a schematic diagram of the system module of the embodiment of the present invention;

[0076] Figure 2 It is a schematic diagram of the specific process of the embodiment of the present invention;

[0077] Figure 3 It is a schematic diagram of the method steps of the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0078] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0079] Please refer to Figure 1 - Figure 2 , the first aspect embodiment of the present invention provides an AI dynamic monitoring and guarding system based on edge computing, including: a data acquisition module, a data analysis module, and a warning module;

[0080] Data acquisition module: used to obtain the security data of the highway toll collection system and the environmental data at the corresponding acquisition time; among them, the security data includes: the operation data of the equipment and the log data of the software;

[0081] Data analysis module: calculates the correlation coefficient of the operation data and the log data, and processes the security data based on the correlation coefficient to obtain the device target data and the software target data; and,

[0082] Calculates the device failure evaluation coefficient and the software failure evaluation coefficient based on the device target data and the software target data; calculates the security evaluation coefficient based on the device failure evaluation coefficient and the software failure evaluation coefficient, and corrects the security evaluation coefficient based on the environmental data to obtain the target security evaluation coefficient;

[0083] Early warning module: judges whether to give an early warning based on the target security evaluation coefficient.

[0084] Obtain the security data of the highway toll collection system and the environmental data at the corresponding acquisition time, including:

[0085] Real-time collect the operation data of the highway toll collection equipment through data sensors; among them, the operation data includes: the temperature, current, voltage, output power, vibration amplitude, noise level and operation time of the equipment, etc.;

[0086] Collect the log data of the highway toll collection software system through data collection equipment; among them, the data collection equipment includes: Fluentd tool or Prometheus tool; the log data includes: API response time, CPU usage rate, memory usage rate, network throughput, log error rate and network latency, etc.;

[0087] Real-time collect the environmental data through data sensors; among them, the environmental data includes: environmental temperature and environmental humidity.

[0088] Calculate the correlation coefficient of the operation data and the log data through calculation, including: calculating the correlation coefficient of the operation data and calculating the correlation coefficient of the log data;

[0089] The method for calculating the correlation coefficient of the operation data is:

[0090] A1: Obtain several groups of operation data and log data from the historical database of the highway toll collection system, and the device status corresponding to each data in the operation data and the software status corresponding to each data in the log data; among them, the device status includes: device failure and device normal; the software status includes: software failure and software normal;

[0091] A2: Mark several groups of operation data as Y i,g ; among them, Yi,g refers to the i-th data in the g-th group of operation data, where g = {1, 2, 3, …, M}, M is the total number of groups, and i = {1, 2, 3, …, N}, N is the total number of data included in the operation data;

[0092] A3: Through the formula P i = G i / g, calculate the probability that each data in the operation data causes equipment failure; where P i refers to the probability that the i-th data in the operation data causes equipment failure, i is the label of each data in the operation data, and G i refers to the total number of times that the i-th data in the operation data causes equipment failure;

[0093] A4: Through the formula XG i = (e ^ (P i ) - e ^ (-P i )) / (e ^ (P i ) + e ^ (-P i )) to obtain the correlation coefficient between each data in the operation data and equipment failure; where XG i refers to the correlation coefficient between the i-th data in the operation data and equipment failure.

[0094] Calculate the correlation coefficient of log data, including:

[0095] B1: Several groups of log data are marked as R j,g ; where R j,g refers to the j-th data in the g-th group of log data, j = {1, 2, 3, …, L}, L is the total number of data included in the log data;

[0096] B2: Through the formula P j = G j / g, calculate the probability that each data in the log data causes software failure; where P j refers to the probability that the j-th data in the log data causes software failure, j is the label of each data in the log data, and G j refers to the total number of times that the j-th data in the log data causes software failure;

[0097] B3: Through the formula XG j = (e ^ (P i ) - e ^ (-P i )) / (e ^ (P i ) + e ^ (-P i )) to obtain the correlation coefficient between each data in the log data and software failure; where XG j refers to the correlation coefficient between the j-th data in the log data and software failure.

[0098] Processing safety data based on the correlation coefficient, including:

[0099] Judging whether the correlation coefficient between each data in the operation data and the equipment failure and the correlation coefficient between each data in the log data and the software failure are greater than the preset correlation coefficient threshold; if yes, retain the corresponding data; if no, remove the corresponding data;

[0100] Mark the retained operation data as equipment target data and the retained log data as software target data.

[0101] For example: The toll collection system of a highway toll station includes hardware devices (such as ETC antennas, lane controllers) and software systems (such as billing services, network communication modules). During the operation of the system, sensors collect the operation data of the equipment in real time, the log collection tool (such as Fluentd) records the log data of the software system, and at the same time, environmental sensors monitor the environmental temperature and humidity; specifically as follows:

[0102] 1. Data collection;

[0103] Operation data:

[0104] Temperature: 42°C;

[0105] Current: 6A;

[0106] Vibration amplitude: 0.8mm;

[0107] Log data:

[0108] API response time: 0.5 seconds;

[0109] CPU usage rate: 80%;

[0110] Memory usage rate: 75%;

[0111] Environmental data:

[0112] Environmental temperature: 42°C;

[0113] Environmental humidity: 75%;

[0114] 2. Calculate the correlation coefficient;

[0115] Obtain several groups of operation data and log data from the historical database, as well as the corresponding equipment status and software status. Calculate the correlation coefficient through the formula:

[0116] Correlation coefficient of operation data:

[0117] Assume the probability that temperature causes equipment failure: 0.7;

[0118] Assume the probability that current causes equipment failure: 0.3;

[0119] Assume the probability that the vibration amplitude causes equipment failure: 0.5;

[0120] Assume that through the formula XG i =(e^(P i )-e^(-P i )) / (e^(P i )+e^(-P i )) Calculate the relevant coefficients:

[0121] The relevant coefficient of temperature: 0.8;

[0122] The relevant coefficient of current: 0.5;

[0123] The relevant coefficient of vibration amplitude: 0.7;

[0124] The relevant coefficient of log data:

[0125] The probability that the API response time causes software failure: 0.6;

[0126] The probability that the CPU usage rate causes software failure: 0.4;

[0127] The probability that the memory usage rate causes software failure: 0.6;

[0128] Through the formula XG j =(e^(P i )-e^(-P i )) / (e^(P i )+e^(-P i )) Calculate the relevant coefficients:

[0129] The relevant coefficient of API response time: 0.7;

[0130] The relevant coefficient of CPU usage rate: 0.4;

[0131] The relevant coefficient of memory usage rate: 0.65;

[0132] 3. Screen the target data;

[0133] Set the preset relevant coefficient threshold to 0.6, and retain the data with the relevant coefficient greater than the threshold:

[0134] Equipment target data: temperature and vibration amplitude;

[0135] Software target data: API response time and memory usage rate.

[0136] The device failure evaluation coefficient and the software failure evaluation coefficient are calculated based on the device target data and the software target data, including: calculating the device failure evaluation coefficient based on the device target data and calculating the software failure evaluation coefficient based on the software target data;

[0137] Calculating the device failure evaluation coefficient based on the device target data includes:

[0138] Extracting the device target data and the corresponding correlation coefficient; Marking the device target data as SM e and marking the correlation coefficient corresponding to the device target data as XG e ; where SM e refers to the e-th data in the device target data, and XG e refers to the correlation coefficient corresponding to the e-th data in the device target data, e = {1, 2, 3,..., Q}, and Q is the total number of data included in the device target data;

[0139] Calculating the device weight factor through the formula x = ∑(1 / XG e ); where x is the device weight factor, ∑ is the summation symbol, and the summation range is e;

[0140] Calculating the weight coefficient corresponding to each data in the device target data through the formula α e = 1 / (XG e ×x); where α e refers to the weight coefficient corresponding to the e-th data in the device target data;

[0141] Calculating the device failure evaluation coefficient through the formula P1 = ∑(α e ×e^((SM e -ZS e ) / ZS e )); where P1 is the device failure evaluation coefficient, ∑ is the summation symbol, the summation range is e, and ZS e refers to the optimal data corresponding to the e-th data in the device target data.

[0142] Calculating the software failure evaluation coefficient based on the software target data includes:

[0143] Extracting the software target data and the corresponding correlation coefficient; Marking the software target data as RM k and marking the correlation coefficient corresponding to the software target data as RX k ; where RM k refers to the k-th data in the software target data, and RX k refers to the correlation coefficient corresponding to the k-th data in the software target data, k = {1, 2, 3,..., V}, and V is the total number of data included in the software target data;

[0144] The software weight factor is calculated by the formula y = ∑(1 / RX k ); where y is the software weight factor, ∑ is the summation symbol, and the summation range is k;

[0145] The weight coefficient corresponding to each data in the software target data is calculated by the formula β k = 1 / (RX k ×y); where β k refers to the weight coefficient corresponding to the kth data in the software target data;

[0146] The software fault evaluation coefficient is calculated by the formula P2 = ∑(β k ×e^((RM k -ZR k ) / ZR k )); where P2 is the software fault evaluation coefficient, ∑ is the summation symbol, the summation range is k, and ZR k refers to the optimal data corresponding to the kth data in the software target data.

[0147] The safety evaluation coefficient is corrected based on the environmental data, including:

[0148] Extracting the device fault evaluation coefficient P1 and the software fault evaluation coefficient P2;

[0149] The safety evaluation coefficient is calculated by the formula P = θ1×P1 + θ2×P2; where P is the safety evaluation coefficient, and θ1 and θ2 are the initial weight coefficients;

[0150] Based on the weather prediction platform, obtain the highest environmental temperature, the lowest environmental temperature, the highest environmental humidity, and the lowest environmental humidity on the same day, and mark them as Tmax, Tmin, Hmax, and Hmin respectively;

[0151] The environmental temperature index is calculated by the formula Tn = (T - Tmin) / (Tmax - Tmin); where Tn is the environmental temperature index and T is the environmental temperature;

[0152] The environmental humidity index is calculated by the formula Hn = (H - Hmin) / (Hmax - Hmin); where Hn is the environmental humidity index and H is the environmental humidity;

[0153] The initial weight coefficient θ1 is corrected based on the environmental temperature index, specifically:

[0154] The corrected weight coefficient of θ1 is calculated by the formula θ1’ = θ1 + f1×Tn - f2×Hn; where θ1’ is the corrected weight coefficient of θ1, f1 is the temperature influence factor, and f2 is the humidity influence factor;

[0155] The initial weight coefficient θ2 is corrected based on the environmental humidity index, specifically as follows:

[0156] The corrected weight coefficient of θ2 is calculated through the formula θ2’ = θ2 - f1×Tn + f2×Hn; where θ2’ is the corrected weight coefficient of θ2;

[0157] By normalizing the corrected weight coefficients θ1’ and θ2’, the target weight coefficients are obtained. The normalization formula is specifically:

[0158] ∈1 = θ1’ / (θ1’ + θ2’) and ∈2 = θ2’ / (θ1’ + θ2’); where ∈1 and ∈2 are the target weight coefficients;

[0159] The initial weight coefficients θ1 and θ2 are updated to ∈1 and ∈2 respectively;

[0160] The target safety assessment coefficient is calculated through the formula P’ = ∈1×P1 + ∈2×P2; where P’ is the target safety assessment coefficient.

[0161] Based on the target safety assessment coefficient, it is judged whether to issue a warning, including:

[0162] Judge whether the target safety assessment coefficient is greater than the preset target safety assessment coefficient threshold; if so, generate a warning message and send it to the client; if not, continue to detect and judge.

[0163] For example: Extract the target data of the device and the corresponding correlation coefficients;

[0164] Suppose XG1 = 0.8; XG2 = 0.7;

[0165] The device weight factor is calculated through the formula x = ∑(1 / XG e )

[0166] Suppose x = 2.68;

[0167] Through the formula α e = 1 / (XG e ×x) calculate the weight coefficient corresponding to each data in the device target data;

[0168] Suppose α1 = 0.46; α2 = 0.54;

[0169] The device failure assessment coefficient is calculated through the formula P1 = ∑(α e ×e^((SM e -ZS e ) / ZS e ))

[0170] Substitute the data to get P1 = 1.429;

[0171] Extract the software target data and the corresponding correlation coefficients;

[0172] Assume RX1 = 0.7 and RX2 = 0.6;

[0173] Calculate the software weight factor through the formula y = ∑(1 / RX k )

[0174] Assume y = 3.1;

[0175] Calculate the weight coefficient corresponding to each data in the software target data through the formula β k = 1 / (RX k ×y)

[0176] Assume β1 = 0.45 and β2 = 0.55;

[0177] Calculate the software fault assessment coefficient through the formula P2 = ∑(β k ×e^((RM k -ZR k ) / ZR k ))

[0178] Substitute the data to get P2 = 1.7685;

[0179] Calculate the safety assessment coefficient through the formula P = θ1×P1 + θ2×P2;

[0180] Obtain the maximum environmental temperature, minimum environmental temperature, maximum environmental humidity and minimum environmental humidity of the day based on the weather prediction platform, and mark them as Tmax, Tmin, Hmax and Hmin respectively;

[0181] Calculate the environmental temperature index through the formula Tn = (T - Tmin) / (Tmax - Tmin);

[0182] Assume Tn = 0.53;

[0183] Calculate the environmental humidity index through the formula Hn = (H - Hmin) / (Hmax - Hmin);

[0184] Assume Hn = 0.83;

[0185] Modify the initial weight coefficient θ1 based on the environmental temperature index, specifically:

[0186] Calculate the modified weight coefficient of θ1 through the formula θ1’ = θ1 + f1×Tn - f2×Hn;

[0187] Substitute the data to get θ1’ = 0.617;

[0188] Modify the initial weight coefficient θ2 based on the environmental humidity index, specifically as follows:

[0189] Calculate the modified weight coefficient of θ2 through the formula θ2’ = θ2 - f1×Tn + f2×Hn;

[0190] Substitute the data to get θ2’ = 0.383;

[0191] Obtain the target weight coefficients by normalizing the modified weight coefficients θ1’ and θ2’, specifically as follows:

[0192] ∈1 = θ1’ / (θ1’ + θ2’) = 0.617, ∈2 = θ2’ / (θ1’ + θ2’) = 0.383;

[0193] Update the initial weight coefficients θ1 and θ2 to ∈1 and ∈2 respectively;

[0194] Calculate the target safety assessment coefficient through the formula P’ = ∈1×P1 + ∈2×P2;

[0195] Substitute the data to get P’ = 1.557;

[0196] Assume that the preset target safety assessment coefficient threshold is 1.5;

[0197] Since P’ = 1.557 > 1.5, the system generates a warning message and sends it to the client.

[0198] This example demonstrates how to comprehensively analyze device operation data, software log data, and environmental data through the method of the present invention, calculate the target safety assessment coefficient, and issue a warning based on the result. This method can effectively predict potential faults and improve the operation efficiency and safety of the system.

[0199] Refer to Figure 3 , the second aspect embodiment of the present invention provides an AI dynamic monitoring and guarding method based on edge computing, including:

[0200] Step 1: Obtain the safety data of the highway toll system and the environmental data at the corresponding collection time;

[0201] Step 2: Calculate the correlation coefficient of the operation data and the log data, and process the safety data based on the correlation coefficient to obtain the device target data and the software target data;

[0202] Step 3: Calculate the device fault assessment coefficient and the software fault assessment coefficient based on the device target data and the software target data;

[0203] Step 4: Calculate the safety evaluation coefficient based on the equipment failure evaluation coefficient and the software failure evaluation coefficient, and correct the safety evaluation coefficient based on the environmental data to obtain the target safety evaluation coefficient;

[0204] Step 5: Determine whether to issue a warning based on the target safety evaluation coefficient.

[0205] Some of the data in the above formula are taken as their numerical values after removing the dimension. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0206] The working principle of the present invention:

[0207] The present invention obtains the safety data of the highway toll collection system and the environmental data at the corresponding collection time; calculates the correlation coefficient of the operation data and the log data, and processes the safety data based on the correlation coefficient to obtain the equipment target data and the software target data; calculates the equipment failure evaluation coefficient and the software failure evaluation coefficient based on the equipment target data and the software target data; calculates the safety evaluation coefficient based on the equipment failure evaluation coefficient and the software failure evaluation coefficient, and corrects the safety evaluation coefficient based on the environmental data to obtain the target safety evaluation coefficient; determines whether to issue a warning based on the target safety evaluation coefficient, solving the technical problems that the prior art can only independently monitor hardware devices or software services and cannot take into account the operating states of both at the same time; and the monitoring system only shows the online or offline state of the device, and manual on-site processing of faults is required, resulting in low response efficiency; potential faults cannot be predicted through data analysis, leading to passive discovery of faults and affecting the operation continuity.

[0208] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An AI dynamic monitoring and protection system based on edge computing, characterized in that: include: Data collection module, data analysis module and early warning module; Data collection module: used to obtain the safety data of the highway toll collection system and the environmental data at the corresponding collection time; the safety data includes: the operation data of the equipment and the log data of the software; Data analysis module: obtains the correlation coefficient between operation data and log data by calculation, and processes the security data based on the correlation coefficient to obtain device target data and software target data; and, The equipment failure assessment coefficient and the software failure assessment coefficient are calculated based on the equipment target data and the software target data; the safety assessment coefficient is calculated based on the equipment failure assessment coefficient and the software failure assessment coefficient, and the safety assessment coefficient is corrected based on the environmental data to obtain the target safety assessment coefficient; Early warning module: Determines whether to issue an early warning based on the target safety assessment coefficient.

2. According to the edge computing-based AI dynamic monitoring and protection system of claim 1, it is characterized in that: The obtaining of the safety data of the highway toll collection system and the environmental data corresponding to the collection time includes: Collect the operation data of highway toll collection equipment in real time through data sensors; Collect log data of the highway toll collection software system through data collection equipment; the data collection equipment includes: Fluentd tool or Prometheus tool; Environmental data is collected in real time through data sensors; the environmental data includes: ambient temperature and ambient humidity.

3. According to the edge computing-based AI dynamic monitoring and protection system of claim 1, it is characterized in that: The step of obtaining the correlation coefficient between the operation data and the log data by calculation includes: calculating the correlation coefficient between the operation data and calculating the correlation coefficient between the log data; The method for calculating the correlation coefficient of the running data is: A1: Obtain several groups of operation data and log data, as well as the device status corresponding to each data in the operation data and the software status corresponding to each data in the log data from the historical database of the highway toll collection system; wherein the device status includes: device failure and device normal; the software status includes: software failure and software normal; A2: Mark several groups of running data as Y i,g ; Among them, Y i,g It refers to the i-th data in the g-th group of running data, g = {1, 2, 3, ..., M}, M is the total number of groups, i = {1, 2, 3, ..., N}, N is the total number of data contained in the running data; A3: Through the formula P i =G i / g calculates the probability of each data in the operating data causing equipment failure; where P i It refers to the probability that the i-th data in the operating data causes equipment failure, i is the label of each data in the operating data, G i It refers to the total number of equipment failures caused by the i-th data in the operating data; A4: By formula XG i =(e^(P i )-e^(-P i )) / (e^(P i )+e^(-P i )) Get the correlation coefficient between each data in the operation data and the equipment failure; Among them, XG i It refers to the correlation coefficient between the i-th data in the operating data and the equipment failure.

4. According to claim 3, an AI dynamic monitoring and protection system based on edge computing is characterized in that: The step of calculating the correlation coefficient of the log data includes: B1: Several groups of log data are marked as R j,g ; Among them, R j,g refers to the jth data in the gth group of log data, j = {1, 2, 3, ..., L}, L is the total number of data contained in the log data; B2: By formula P j =G j / g calculates the probability of each data in the log data causing software failure; where P j It refers to the probability that the jth data in the log data causes software failure, j is the label of each data in the log data, G j It refers to the total number of software failures caused by the jth data in the log data; B3: By formula XG j =(e^(P i )-e^(-P i )) / (e^(P i )+e^(-P i )) Get the correlation coefficient between each data in the log data and the software failure; where XG j It refers to the correlation coefficient between the jth data in the log data and the software failure.

5. According to the edge computing-based AI dynamic monitoring and protection system of claim 1, it is characterized in that: The processing of the security data based on the correlation coefficient includes: Determine whether the correlation coefficient between each data in the operation data and the equipment failure and the correlation coefficient between each data in the log data and the software failure is greater than a preset correlation coefficient threshold; if yes, retain the corresponding data; if no, remove the corresponding data; The retained operating data is marked as device target data, and the retained log data is marked as software target data.

6. According to the edge computing-based AI dynamic monitoring and protection system of claim 1, it is characterized in that: The step of calculating the device fault assessment coefficient and the software fault assessment coefficient based on the device target data and the software target data comprises: calculating the device fault assessment coefficient based on the device target data and calculating the software fault assessment coefficient based on the software target data; The device failure assessment coefficient is calculated based on the device target data, including: Extract device target data and corresponding correlation coefficients; mark the device target data as SM e , mark the correlation coefficient corresponding to the device target data as XG e ; Among them, SM e Refers to the eth data in the device target data, XG e It refers to the correlation coefficient corresponding to the e-th data in the device target data, e = {1, 2, 3, …, Q}, Q is the total number of data contained in the device target data; By the formula x = ∑(1 / XG e ) calculates the device weight factor; where x is the device weight factor, ∑ is the summation factor, and the summation range is e; Through the formula α e =1 / (XG e ×x) to calculate the weight coefficient corresponding to each data in the device target data; where α e It refers to the weight coefficient corresponding to the e-th data in the device target data; By the formula P1 = ∑(α e ×e^((SM e -ZS e ) / ZS e )) Calculate the equipment failure assessment coefficient; where P1 refers to the equipment failure assessment coefficient, ∑ is the summation, the summation range is e, ZS e It refers to the optimal data corresponding to the e-th data in the device target data.

7. According to claim 6, an AI dynamic monitoring and protection system based on edge computing is characterized in that: The calculating of the software fault assessment coefficient based on the software target data includes: Extract software target data and corresponding correlation coefficients; mark the software target data as RM k , mark the correlation coefficient corresponding to the software target data as RX k ; Among them, RM k Refers to the kth data in the software target data, RX k It refers to the correlation coefficient corresponding to the kth data in the software target data, k = {1, 2, 3, …, V}, V is the total number of data contained in the software target data; By the formula y = ∑(1 / RX k ) calculates the software weight factor; wherein y is the software weight factor, ∑ is the summation, and the summation range is k; Through the formula β k =1 / (RX k ×y) to calculate the weight coefficient corresponding to each data in the software target data; where β k It refers to the weight coefficient corresponding to the kth data in the software target data; By the formula P2 = ∑(β k ×e^((RM k -ZR k ) / ZR k )) Calculate the software fault assessment coefficient; where P2 refers to the software fault assessment coefficient, ∑ is the summation, the summation range is k, ZR k It refers to the optimal data corresponding to the kth data in the software target data.

8. According to claim 1, an AI dynamic monitoring and protection system based on edge computing is characterized in that: The correction of the safety assessment coefficient based on environmental data includes: Extracting equipment fault assessment coefficient P1 and software fault assessment coefficient P2; The safety assessment coefficient is calculated by the formula P = θ1 × P1 + θ2 × P2; where P is the safety assessment coefficient, θ1 and θ2 are the initial weight coefficients; Based on the weather forecast platform, the highest ambient temperature, lowest ambient temperature, highest ambient humidity and lowest ambient humidity of the day are obtained and marked as Tmax, Tmin, Hmax and Hmin respectively; The ambient temperature index is calculated by the formula Tn=(T-Tmin) / (Tmax-Tmin); wherein Tn is the ambient temperature index and T is the ambient temperature; The ambient humidity index is calculated by the formula Hn = (H-Hmin) / (Hmax-Hmin); wherein Hn is the ambient humidity index and H is the ambient humidity; The initial weight coefficient θ1 is corrected based on the ambient temperature index, specifically: The weight coefficient of θ1 after correction is calculated by the formula θ1'=θ1+f1×Tn-f2×Hn; wherein θ1' is the weight coefficient of θ1 after correction, f1 is the temperature influence factor, and f2 is the humidity influence factor; The initial weight coefficient θ2 is corrected based on the ambient humidity index, specifically: The weight coefficient of θ2 after correction is calculated by the formula θ2'=θ2-f1×Tn+f2×Hn; wherein θ2' is the weight coefficient of θ2 after correction; By normalizing the corrected weight coefficients θ1' and θ2', the target weight coefficient is obtained. The normalization formula is as follows: ∈1=θ1' / (θ1'+θ2') and ∈2=θ2' / (θ1'+θ2'); where ∈1 and ∈2 are target weight coefficients; Update the initial weight coefficients θ1 and θ2 to ∈1 and ∈2 respectively; The target safety assessment coefficient is calculated by the formula P'=∈1×P1+∈2×P2; wherein P' is the target safety assessment coefficient.

9. The AI ​​dynamic monitoring and protection system based on edge computing according to claim 1 is characterized in that: The determining whether to issue an early warning based on the target safety assessment coefficient includes: Determine whether the target security assessment coefficient is greater than the preset target security assessment coefficient threshold; if yes, generate an early warning message and send it to the client; if not, continue to detect and judge.

10. An AI dynamic monitoring and guarding method based on edge computing, applied to an AI dynamic monitoring and guarding system based on edge computing according to any one of claims 1 to 9, characterized in that: include: Step 1: Obtain the safety data of the highway toll collection system and the environmental data at the corresponding collection time; Step 2: Obtain the correlation coefficient between the operation data and the log data by calculation, and process the security data based on the correlation coefficient to obtain the device target data and the software target data; Step 3: Calculate the equipment fault assessment coefficient and the software fault assessment coefficient based on the equipment target data and the software target data; Step 4: Calculate the safety assessment coefficient based on the equipment fault assessment coefficient and the software fault assessment coefficient, and modify the safety assessment coefficient based on the environmental data to obtain the target safety assessment coefficient; Step 5: Determine whether to issue an early warning based on the target safety assessment coefficient.

Images