Circuit controller fault prediction method and system based on machine learning

Abstract

The invention discloses a circuit controller fault prediction method and system based on machine learning, and relates to the technical field of fault prediction.The circuit controller fault prediction method comprises the steps of 1, circuit controller regular detection, 2, fault prediction analysis and 3, circuit controller fault prediction localization. The traffic signal control machine can be detected at midnight, the workload of workers is reduced, the influence on traffic is reduced, and whether electric leakage, line power supply faults and other possible problems exist in a circuit controller or not is accurately judged by analyzing various data between each button and each related device, so that timely early warning is carried out, and the traffic signal control machine is prevented from being damaged. The fault occurrence rate of the circuit controller of the traffic signal control machine is reduced, early warning is carried out on a person in charge, the person in charge is reminded to maintain the lossy traffic signal control machine in time, and the situation that the traffic signal control machine cannot be used due to serious faults is prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a circuit controller fault prediction method and system based on machine learning. Background Art

[0002] The stable operation of traffic signal controllers is directly related to the efficiency and safety of urban road traffic. The circuit controller is a key component of the traffic signal controller, responsible for functions such as the timing control of traffic lights. However, due to long-term exposure to complex environments, the circuit controller is prone to various types of losses. If these losses are not maintained in time, they can easily lead to various types of failures, thereby reducing the operating efficiency of the traffic signal controller. Therefore, it is very necessary to predict the faults of the circuit controller of the traffic signal controller.

[0003] Prior art, such as the invention patent application with publication number CN118797401A, discloses a storage device and operating method for predicting faults using machine learning. The method includes receiving risk data to learn patterns in the data, and outputting an anomaly score for the received data based on the learned patterns. Prior art, such as the invention patent application with publication number CN115516254A, discloses a system and method for controlling variable refrigerant flow systems and equipment using artificial intelligence models. The method includes the processing circuit being configured to use a machine learning model to analyze operating data of the HVAC equipment to predict a variable state or condition of the oil used in the HVAC equipment. The processing circuit is configured to identify insufficient oil based on the variable state or condition of the oil. The processing circuit is configured to automatically initiate corrective action in response to identifying insufficient oil.

[0004] As can be seen from the above scheme, the current circuit controller fault prediction method mostly adopts regular manual inspection. If the circuit controller of the traffic signal controller needs to be actually debugged, it is necessary to observe the feedback of the traffic light. If there are vehicles driving during the day, the debugging of the traffic light is likely to affect traffic. If there are no vehicles in the middle of the night, working in the middle of the night is likely to increase the manual burden. In addition, some traffic signal controllers are located in remote areas, and it is often impossible to predict the faults of the circuit controllers of these traffic signal controllers in a timely manner, thereby increasing the failure rate of the traffic signal controllers. Summary of the Invention

[0005] The purpose of the present invention is to provide a circuit controller fault prediction method and system based on machine learning, which solves the problems existing in the background technology.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions: The first aspect of the present invention provides a circuit controller fault prediction method based on machine learning, including: Step 1. Regular detection of circuit controller: At the target monitoring time point, an intelligent robot is dispatched to detect the relevant equipment of the circuit controller belonging to the traffic signal controller, and the relevant equipment includes: buttons, internal prompt lights and external related signal lights.

[0007] Step 2. Fault prediction analysis: Evaluate the damage coefficient of pressing each button of the circuit controller of the traffic signal controller, analyze the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller of the traffic signal controller is pressed, and calculate the line fault warning coefficient between each button of the circuit controller of the traffic signal controller and other related equipment.

[0008] Step 3. Circuit controller fault prediction and location: Analyze the estimated fault locations of the circuit controllers belonging to the traffic signal controller and send them to the person in charge of traffic signal controller management.

[0009] The second aspect of the present invention provides a fault prediction system for executing the machine learning-based circuit controller fault prediction method, including: a circuit controller periodic detection module, which is used to dispatch an intelligent robot to detect the relevant equipment of the circuit controller belonging to the traffic signal controller at the target monitoring time point, and the relevant equipment includes: buttons, internal prompt lights and external related signal lights.

[0010] The fault prediction analysis module is used to evaluate the pressing damage coefficient of each button of the circuit controller belonging to the traffic signal control machine, analyze the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal control machine is pressed, and calculate the line fault warning coefficient between each button of the circuit controller belonging to the traffic signal control machine and other related equipment.

[0011] The circuit controller fault prediction and location module is used to analyze the estimated fault locations of the circuit controllers belonging to the traffic signal controller and send them to the person in charge of traffic signal controller management.

[0012] The beneficial effects of the present invention are as follows: (1) Step 1 of the present invention performs regular inspections on the circuit controller, and the intelligent robot automatically presses buttons and collects multimodal data, so that the traffic signal controller can be inspected in the middle of the night, which not only reduces the workload of manual labor but also reduces the impact on traffic.

[0013] (2) Step 2 of the present invention is to analyze various data between each button and each related device to accurately determine whether there are potential problems such as leakage and line power supply failure in the circuit controller, thereby providing timely warnings and reducing the incidence of circuit controller failures in traffic signal controllers.

[0014] (3) Step 3 of the present invention predicts and locates circuit controller faults, and issues an early warning to the person in charge, reminding him to promptly maintain the damaged traffic signal controller, thereby preventing serious faults from causing the controller to be unusable. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 Schematic diagram of the method of the present invention.

[0017] Figure 2 Schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Reference Figure 1 As shown, the first aspect of the present invention provides a circuit controller fault prediction method based on machine learning, including: 1. A circuit controller fault prediction method based on machine learning, characterized in that it includes: Step 1. Regular detection of the circuit controller: At the target monitoring time point, an intelligent robot is dispatched to detect the relevant equipment of the circuit controller belonging to the traffic signal controller, and the relevant equipment includes: various buttons, various internal prompt lights and various external related signal lights.

[0020] It should be noted that the target monitoring time point is a time point when there are no vehicles at the intersection managed by the traffic signal controller.

[0021] In a specific embodiment of the present invention, the detection of the relevant devices of the circuit controller belonging to the traffic signal controller is carried out by a specific detection method: obtaining a surface image of the circuit controller belonging to the traffic signal controller, and obtaining the location of each button of the circuit controller belonging to the traffic signal controller through image recognition technology.

[0022] It should be noted that the existing image recognition technology is relatively mature. Through the existing image recognition technology and based on the surface image of the circuit controller of the traffic signal controller, the location of each button of the circuit controller of the traffic signal controller can be obtained.

[0023] The intelligent robot's end execution part is used to press the positions of the buttons of the circuit controller belonging to the traffic signal controller in sequence, and obtain the pressing force characteristic value of the circuit controller belonging to the traffic signal controller after each button is pressed, the success time point of each external related signal light and the image at each monitoring time point, the success time point of each internal warning light and the image at each monitoring time point, and obtain the voltage of each leakage detection point of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed.

[0024] It should be noted that the pressing force characteristic value is the force applied by the end-executing part of the intelligent robot to each button of the circuit controller belonging to the traffic signal control machine when the button is successfully pressed. The end-executing part of the intelligent robot will gradually increase the applied force and monitor the output voltage of the button when pressing, and obtain the output voltage threshold from the local database. If the output voltage of the button is greater than the output voltage threshold during the process of increasing the applied force of the end-executing part of the intelligent robot, then the applied force of the end-executing part of the intelligent robot at this moment is used as the pressing force characteristic value.

[0025] It should also be noted that the success time point is the time point when the system inside the circuit controller itself believes that the adjustment has been completed after receiving current feedback.

[0026] In a specific embodiment, the surface image of the circuit controller of the traffic signal controller is obtained by a specific method of photographing the circuit controller of the traffic signal controller through a camera of an intelligent robot, thereby obtaining the surface image of the circuit controller of the traffic signal controller.

[0027] In a specific embodiment, the pressing force characteristic value of the circuit controller belonging to the traffic signal controller after each button is pressed, the success time point of each external related signal light and the image at each monitoring time point, the success time point of each internal warning light and the image at each monitoring time point, and the voltage of each leakage detection point of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed are obtained. The specific acquisition method is: obtaining on the intelligent robot terminal, the mounted camera and the voltage sensor.

[0028] It should be noted that the local database is used to store the output voltage threshold, the target usage time of the traffic signal controller, the pressing force characteristic value of each button of each other traffic signal controller at each usage time, the safe voltage of the leakage detection point, the appropriate number of target pixels, the allowed feedback time, the target grayscale value range of each pixel point in the image of each external related traffic light after each button is pressed, the pressing damage coefficient threshold, the leakage warning hazard coefficient threshold, and the line fault warning coefficient threshold.

[0029] Step 1 of the present invention is to regularly inspect the circuit controller. By automatically pressing buttons and collecting multimodal data through an intelligent robot, the traffic signal controller can be inspected in the middle of the night, which not only reduces the manual workload but also reduces the impact on traffic.

[0030] Step 2. Fault prediction analysis: Evaluate the damage coefficient of pressing each button of the circuit controller of the traffic signal controller, analyze the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller of the traffic signal controller is pressed, and calculate the line fault warning coefficient between each button of the circuit controller of the traffic signal controller and other related equipment.

[0031] It should be noted that the other related equipment include: internal warning lights and external related signal lights.

[0032] In a specific embodiment of the present invention, the method for evaluating the pressing damage coefficient of each button of the circuit controller of the traffic signal controller is as follows: obtaining the target usage time of the traffic signal controller and the pressing force characteristic value of each button of each other traffic signal controller at each usage time from the local database, and mapping the pressing force characteristic value of each button of each other traffic signal controller at the target usage time to obtain the pressing force characteristic value of each button of each other traffic signal controller at the target usage time. , where i represents the number of each other traffic signal controller, , j is a positive integer greater than 2, n represents the number of each button, , m is a positive integer greater than 2.

[0033] It should be noted that the characteristic values ​​of the pressing force of each button of each other traffic signal controller at each usage duration are obtained by conducting long-term tests on other traffic signal controllers of the same model.

[0034] According to the characteristic value of the pressing force of each button of the traffic signal controller , calculate the pressing damage coefficient of each button of the circuit controller of the traffic signal controller , where j represents the number of other traffic signal controllers.

[0035] In a specific embodiment of the present invention, the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal controller is analyzed, and the specific analysis method is: obtaining the safe voltage A of the leakage detection point from the local database.

[0036] According to the voltage of each leakage detection point at each monitoring time point after each button is pressed by the circuit controller of the traffic signal controller , where x represents the number of each monitoring time point, , y is a positive integer greater than 2, p represents the number of each leakage detection point, calculate the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal control machine is pressed , where e is a natural constant.

[0037] In a specific embodiment of the present invention, the circuit fault warning coefficient between each button of the circuit controller of the traffic signal controller and each other related equipment is calculated by a specific analysis method: based on the successful time point of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller and the image at each monitoring time point, the comprehensive hazard coefficient of the feedback time between each button of the circuit controller of the traffic signal controller and each external related signal light is evaluated. , and calculate the image change stability coefficient of each button of the circuit controller of the traffic signal control machine and each external related signal light , where r represents the number of each external related signal light, , s is a positive integer greater than 2.

[0038] Calculate the line fault warning coefficient between each button of the circuit controller of the traffic signal control machine and each external related signal light .

[0039] Similarly, calculate the line fault warning coefficient between each button and each internal indicator light of the circuit controller belonging to the traffic signal control machine.

[0040] In summary, the line fault warning coefficients between each button of the circuit controller of the traffic signal control machine and other related equipment are summarized.

[0041] In a specific embodiment of the present invention, the comprehensive hazard coefficient of the feedback duration of each button of the circuit controller belonging to the traffic signal controller and each external related signal light is evaluated by a specific evaluation method: based on the image of each external related signal light at each monitoring time point after each button of the circuit controller belonging to the traffic signal controller is pressed, the grayscale value of each pixel point of the image of each external related signal light at each monitoring time point after each button of the circuit controller belonging to the traffic signal controller is extracted, and based on this, the target number of pixels of each external related signal light at each monitoring time point after each button of the circuit controller belonging to the traffic signal controller is counted.

[0042] Obtain the appropriate target pixel number from the local database. If the target pixel number of a certain external related signal light at each monitoring time point after a certain button is pressed by the circuit controller of the traffic signal controller is less than the appropriate target pixel number, then the comprehensive hazard coefficient of the feedback time of the button of the circuit controller of the traffic signal controller and the external related signal light is recorded as e. Otherwise, the actual feedback time of each external related signal light after each button of the circuit controller of the traffic signal controller is calculated. .

[0043] According to the success time point of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller, and according to the pressing time point of each button of the circuit controller of the traffic signal controller, the system feedback time of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller is calculated. .

[0044] Obtain the allowed feedback time D from the local database and calculate the comprehensive hazard coefficient of the feedback time of each button of the circuit controller of the traffic signal controller and each external related signal light. .

[0045] In summary, the comprehensive hazard coefficient of the feedback time of each button of the circuit controller belonging to the traffic signal control machine and each external related signal light is summarized.

[0046] In a specific embodiment, the target pixel number of each external related signal light of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed is counted by a specific statistical method as follows: a target grayscale value interval of each pixel of the image of each external related signal light after each button is pressed is obtained from a local database; if the grayscale value of a certain pixel of the image of a certain external related signal light at a certain monitoring time point after a certain button is pressed by the circuit controller belonging to the traffic signal controller is within the target grayscale value interval, the pixel is marked as a target pixel, thereby screening the target pixel points of the image of each external related signal light of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed, and counting the target pixel number of each external related signal light of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed.

[0047] In a specific embodiment of the present invention, the actual feedback duration of each external related signal light of the circuit controller belonging to the traffic signal controller after each button is pressed is calculated, and its specific calculation method is: obtaining the pressing time points of each button of the circuit controller belonging to the traffic signal controller from the intelligent robot terminal, if the number of target pixels of a certain external related signal light of the circuit controller belonging to the traffic signal controller at a certain monitoring time point after a certain button is pressed is greater than the appropriate target pixel number, and the number of target pixels at the first ranked monitoring time point is greater than the appropriate target pixel number, then the monitoring time point is marked as the actual monitoring time point, thereby screening the actual monitoring time points of each external related signal light of the circuit controller belonging to the traffic signal controller after each button is pressed, and according to the pressing time points of each button of the circuit controller belonging to the traffic signal controller, the actual feedback duration of each external related signal light of the circuit controller belonging to the traffic signal controller after each button is pressed is calculated.

[0048] In a specific embodiment of the present invention, the image change stability coefficient of each button of the circuit controller belonging to the traffic signal controller and each external related signal light is calculated by the following specific calculation method: if the number of target pixel points of a certain external related signal light at each monitoring time point after a certain button of the circuit controller belonging to the traffic signal controller is less than the appropriate target pixel point number, then the image change stability coefficient of the button of the circuit controller belonging to the traffic signal controller and the external related signal light is recorded as 0; otherwise, the image of each external related signal light at each actual target time point after each button of the circuit controller belonging to the traffic signal controller is extracted, and the grayscale value of each pixel point of the image of each external related signal light at each actual target time point after each button of the circuit controller belonging to the traffic signal controller is extracted. , where t represents the actual target time point, , w is a positive integer greater than 2, h represents the number of each pixel, , k is a positive integer greater than 2.

[0049] Calculate the image change stability coefficient of each button of the circuit controller of the traffic signal control machine and each external related signal light .

[0050] It should be noted that the images of the external related signal lights at the actual target time points after the buttons of the circuit controller of the traffic signal control machine are images of the monitoring time points after the actual monitoring time points.

[0051] Step 2 of the present invention is fault prediction analysis. By analyzing the data between each button and each related device, it is accurately determined whether there are potential problems such as leakage and line power supply failure in the circuit controller, thereby providing timely warnings and reducing the incidence of circuit controller failures in traffic signal controllers.

[0052] Step 3. Circuit controller fault prediction and location: Analyze the estimated fault locations of the circuit controllers belonging to the traffic signal controller and send them to the person in charge of traffic signal controller management.

[0053] In a specific embodiment of the present invention, the analysis of the estimated fault locations of the circuit controller belonging to the traffic signal controller is carried out by obtaining a press damage coefficient threshold from a local database; if the press damage coefficient of a button of the circuit controller belonging to the traffic signal controller is greater than the press damage coefficient threshold, the button is marked as a button with an estimated fault; if the pressing force characteristic value of the circuit controller belonging to the traffic signal controller after pressing a button cannot be obtained, the button is marked as a button with an estimated fault, thereby screening the estimated fault buttons of the circuit controller belonging to the traffic signal controller.

[0054] A leakage warning hazard coefficient threshold is obtained from the local database. If the leakage warning hazard coefficient of a leakage detection point when a button of the circuit controller belonging to the traffic signal controller is pressed is greater than the leakage warning hazard coefficient threshold, the line between the button and the leakage monitoring point is marked as an estimated leakage fault line, thereby screening the estimated leakage fault lines of the circuit controller belonging to the traffic signal controller.

[0055] A line fault warning coefficient threshold is obtained from the local database. If the line fault warning coefficient between a button of the circuit controller belonging to the traffic signal controller and some other related equipment is greater than the line fault warning coefficient threshold, the power supply line between the button and the other related equipment is marked as an estimated power supply abnormality line, thereby screening the estimated power supply abnormality lines of the circuit controller belonging to the traffic signal controller.

[0056] Each predicted fault button of the circuit controller belonging to the traffic signal controller, each predicted leakage fault line and each predicted power supply abnormality line are uniformly marked as each predicted fault location.

[0057] Step 3 of the present invention predicts and locates circuit controller faults, and issues an early warning to the person in charge to remind him to perform timely maintenance on the damaged traffic signal controller to prevent serious faults from causing it to be unusable.

[0058] Reference Figure 2 As shown, the second aspect of the present invention provides a fault prediction system for executing the circuit controller fault prediction method based on machine learning, including: a circuit controller periodic detection module, a fault prediction analysis module, a circuit controller fault prediction and positioning module and a local database.

[0059] It should be noted that the circuit controller periodic detection module is connected to the fault prediction and analysis module, the fault prediction and analysis module is connected to the circuit controller fault prediction and positioning module, and the local database is connected to the fault prediction and analysis module and the circuit controller fault prediction and positioning module.

[0060] The circuit controller periodic detection module is used to dispatch an intelligent robot to detect the relevant equipment of the circuit controller belonging to the traffic signal controller at the target monitoring time point. The relevant equipment includes: buttons, internal prompt lights and external related signal lights.

[0061] The fault prediction and analysis module is used to evaluate the damage coefficient of pressing each button of the circuit controller belonging to the traffic signal controller, analyze the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal controller is pressed, and calculate the line fault warning coefficient between each button of the circuit controller belonging to the traffic signal controller and other related equipment.

[0062] The circuit controller fault prediction and location module is used to analyze each estimated fault location of the circuit controller belonging to the traffic signal controller and send the estimated fault location to the person in charge of traffic signal controller management.

[0063] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. A circuit controller fault prediction method based on machine learning, characterized in that: include: Step 1. Regular inspection of the circuit controller: At the target monitoring time, an intelligent robot is dispatched to inspect the relevant devices of the circuit controller of the traffic signal controller, including buttons, internal indicator lights, and external signal lights. Step 2. Fault Prediction Analysis: Evaluate the damage coefficient of pressing each button on the traffic signal controller's circuit controller, analyze the leakage warning hazard coefficient of each leakage detection point when pressing each button on the traffic signal controller's circuit controller, and calculate the line fault warning coefficient between each button on the traffic signal controller's circuit controller and other related equipment. Step 3. Circuit controller fault prediction and location: Analyze the estimated fault locations of the circuit controllers belonging to the traffic signal controller and send them to the person in charge of traffic signal controller management.

2. The circuit controller fault prediction method based on machine learning according to claim 1, characterized in that: The specific detection method for detecting the relevant devices of the circuit controller of the traffic signal controller is as follows: Obtaining a surface image of a circuit controller of a traffic signal control machine, and obtaining the position of each button of the circuit controller of the traffic signal control machine through image recognition technology; The intelligent robot's end execution part is used to press the positions of the buttons of the circuit controller belonging to the traffic signal controller in sequence, and obtain the pressing force characteristic value of the circuit controller belonging to the traffic signal controller after each button is pressed, the success time point of each external related signal light and the image at each monitoring time point, the success time point of each internal warning light and the image at each monitoring time point, and obtain the voltage of each leakage detection point of the circuit controller belonging to the traffic signal controller at each monitoring time point after each button is pressed.

3. The circuit controller fault prediction method based on machine learning according to claim 2, characterized in that: The specific evaluation method for evaluating the pressing damage coefficient of each button of the circuit controller of the traffic signal controller is as follows: Obtain the target usage time of the traffic signal controller and the characteristic value of the pressing force of each button of each other traffic signal controller under each usage time from the local database, and map the characteristic value of the pressing force of each button of each other traffic signal controller under the target usage time. , where i represents the number of each other traffic signal controller, , j is a positive integer greater than 2, n represents the number of each button, , m is a positive integer greater than 2; According to the characteristic value of the pressing force of each button of the traffic signal controller , calculate the pressing damage coefficient of each button of the circuit controller of the traffic signal controller , where j represents the number of other traffic signal controllers.

4. The circuit controller fault prediction method based on machine learning according to claim 3, characterized in that: The specific analysis method for analyzing the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal control machine is pressed is as follows: Obtain the safe voltage A of the leakage detection point from the local database; According to the voltage of each leakage detection point at each monitoring time point after each button is pressed by the circuit controller of the traffic signal controller , where x represents the number of each monitoring time point, , y is a positive integer greater than 2, p represents the number of each leakage detection point, calculate the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller belonging to the traffic signal control machine is pressed , where e is a natural constant.

5. The circuit controller fault prediction method based on machine learning according to claim 3, characterized in that: The specific analysis method for calculating the line fault warning coefficient between each button of the circuit controller of the traffic signal controller and other related equipment is as follows: Based on the success time points of each external related signal light after each button of the circuit controller of the traffic signal controller is pressed and the images at each monitoring time point, the comprehensive hazard coefficient of the feedback time length of each button of the circuit controller of the traffic signal controller and each external related signal light is evaluated. , and calculate the image change stability coefficient of each button of the circuit controller of the traffic signal control machine and each external related signal light , where r represents the number of each external related signal light, , s is a positive integer greater than 2; Calculate the line fault warning coefficient between each button of the circuit controller of the traffic signal control machine and each external related signal light ; Similarly, calculate the line fault warning coefficient between each button and each internal indicator light of the circuit controller of the traffic signal control machine; In summary, the line fault warning coefficients between each button of the circuit controller of the traffic signal control machine and other related equipment are summarized.

6. The circuit controller fault prediction method based on machine learning according to claim 5, characterized in that: The specific evaluation method for evaluating the comprehensive hazard coefficient of the feedback time of each button of the circuit controller of the traffic signal controller and each external related signal light is as follows: Extracting the grayscale value of each pixel of the image of each external related signal light at each monitoring time point after each button of the circuit controller of the traffic signal controller is pressed, and counting the number of target pixels of each external related signal light at each monitoring time point after each button of the circuit controller of the traffic signal controller is pressed based on the grayscale value of each pixel of the image of each external related signal light at each monitoring time point after each button of the circuit controller of the traffic signal controller is pressed; Obtain the appropriate target pixel number from the local database. If the target pixel number of a certain external related signal light at each monitoring time point after a certain button is pressed by the circuit controller of the traffic signal controller is less than the appropriate target pixel number, then the comprehensive hazard coefficient of the feedback time of the button of the circuit controller of the traffic signal controller and the external related signal light is recorded as e. Otherwise, the actual feedback time of each external related signal light after each button of the circuit controller of the traffic signal controller is calculated. ; According to the success time point of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller, and according to the pressing time point of each button of the circuit controller of the traffic signal controller, the system feedback time of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller is calculated. ; Obtain the allowed feedback time D from the local database and calculate the comprehensive hazard coefficient of the feedback time of each button of the circuit controller of the traffic signal controller and each external related signal light. ; In summary, the comprehensive hazard coefficient of the feedback time of each button of the circuit controller belonging to the traffic signal control machine and each external related signal light is summarized.

7. The circuit controller fault prediction method based on machine learning according to claim 6, characterized in that: The calculation method for obtaining the actual feedback duration of each external related signal light after each button is pressed by the circuit controller of the traffic signal controller is as follows: The pressing time points of each button of the circuit controller belonging to the traffic signal controller are obtained from the intelligent robot terminal. If the number of target pixels of a certain external related signal light of the circuit controller belonging to the traffic signal controller at a certain monitoring time point after a certain button is pressed is greater than the appropriate target pixel number, and the number of target pixels at the first ranked monitoring time point is greater than the appropriate target pixel number, then the monitoring time point is marked as the actual monitoring time point, thereby screening the actual monitoring time points of each external related signal light of the circuit controller belonging to the traffic signal controller after each button is pressed, and according to the pressing time points of each button of the circuit controller belonging to the traffic signal controller, the actual feedback duration of each external related signal light of the circuit controller belonging to the traffic signal controller after each button is pressed is calculated.

8. The circuit controller fault prediction method based on machine learning according to claim 7, characterized in that: The specific calculation method for calculating the image change stability coefficient of each button of the circuit controller of the traffic signal control machine and each external related signal light is as follows: If the number of target pixels of a certain external related signal light at each monitoring time point after a certain button of the circuit controller belonging to the traffic signal controller is less than the appropriate target pixel number, the image change stability coefficient of the button and the external related signal light of the circuit controller belonging to the traffic signal controller is recorded as 0. Otherwise, the image of each external related signal light at each actual target time point after each button of the circuit controller belonging to the traffic signal controller is extracted, and the grayscale value of each pixel of the image of each external related signal light at each actual target time point after each button of the circuit controller belonging to the traffic signal controller is extracted. , where t represents the actual target time point, , w is a positive integer greater than 2, h represents the number of each pixel, , k is a positive integer greater than 2; Calculate the image change stability coefficient of each button of the circuit controller of the traffic signal control machine and each external related signal light .

9. The circuit controller fault prediction method based on machine learning according to claim 5, characterized in that: The specific analysis method for analyzing the estimated fault locations of the circuit controllers belonging to the traffic signal controller is as follows: A pressure damage coefficient threshold is obtained from a local database. If the pressure damage coefficient of a button on a circuit controller belonging to a traffic signal controller is greater than the pressure damage coefficient threshold, the button is marked as a button with a predicted fault. If the characteristic value of the pressure force after pressing a button on the circuit controller belonging to the traffic signal controller cannot be obtained, the button is marked as a button with a predicted fault, thereby screening the predicted fault buttons of the circuit controller belonging to the traffic signal controller. Obtaining a leakage warning hazard coefficient threshold from a local database, if the leakage warning hazard coefficient of a leakage detection point when a button of a circuit controller belonging to a traffic signal controller is pressed is greater than the leakage warning hazard coefficient threshold, then marking the line between the button and the leakage detection point as an estimated leakage fault line, thereby screening the estimated leakage fault lines of the circuit controller belonging to the traffic signal controller; Obtaining a line fault warning coefficient threshold from a local database; if the line fault warning coefficient between a button of a circuit controller belonging to a traffic signal controller and a related device is greater than the line fault warning coefficient threshold, marking the power supply line between the button and the related device as an estimated abnormal power supply line, thereby screening the estimated abnormal power supply lines of the circuit controller belonging to the traffic signal controller; Each predicted fault button of the circuit controller belonging to the traffic signal controller, each predicted leakage fault line and each predicted power supply abnormality line are uniformly marked as each predicted fault location.

10. A fault prediction system for executing the circuit controller fault prediction method based on machine learning according to any one of claims 1 to 9, characterized in that: include: The circuit controller periodic detection module is used to dispatch an intelligent robot to detect the relevant equipment of the circuit controller of the traffic signal controller at the target monitoring time point, including: buttons, internal indicator lights and external related signal lights; A fault prediction analysis module is used to evaluate the damage coefficient of pressing each button of the circuit controller of the traffic signal controller, analyze the leakage warning hazard coefficient of each leakage detection point when each button of the circuit controller of the traffic signal controller is pressed, and calculate the line fault warning coefficient between each button of the circuit controller of the traffic signal controller and other related equipment; The circuit controller fault prediction and location module is used to analyze the estimated fault locations of the circuit controllers belonging to the traffic signal controller and send them to the person in charge of traffic signal controller management.

Images