A highway monitoring system maintenance method and system

Abstract

The present invention discloses a highway monitoring system maintenance method and system, and relates to the field of computer technology. A highway monitoring system maintenance system includes: a target system maintenance project acquisition module, a risk prediction module for projects to be maintained, and a priority assessment module for projects to be maintained. The present invention can quickly identify and prioritize high-risk monitoring projects through automated risk assessment and priority evaluation models, significantly improve the efficiency of maintenance work, and ensure that the monitoring system is always in the best operating state, providing strong technical support for highway traffic safety, and can adapt to different traffic conditions and environmental changes, ensuring that effective monitoring system maintenance can be achieved under various circumstances.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a highway monitoring system maintenance method and system. Background Art

[0002] Highway monitoring systems are essential to ensure road traffic safety and efficient operation. Traditional maintenance methods use the principle of "first come, first served" to repair faults when they occur, but this method may result in some emergency or critical faults not being handled in time. Since the repair time and urgency of different faults vary, and faults may be scattered in different locations, the allocation of maintenance resources needs to be more accurate and efficient to ensure that highway monitoring systems can continue to work stably. Summary of the invention

[0003] The present invention aims to provide a highway monitoring system maintenance method and system to accurately allocate maintenance resources.

[0004] A highway monitoring system maintenance method, comprising:

[0005] S1. Get the current target monitoring system to be maintained list; the target monitoring system to be maintained list contains M items to be maintained P m ,m=1,2,…,M;Based on the target monitoring system project failure risk model, the target monitoring system to-be-maintained list is screened to obtain the target monitoring system to-be-maintained project risk list, which includes the to-be-maintained projects with monitoring risks; The target monitoring system project failure risk model is used to select the to-be-maintained projects P m Perform risk value screening, and screen high-risk maintenance items for priority maintenance; if the risk list of the target monitoring system's maintenance items is not empty, remove the maintenance items in the risk list of the target monitoring system's maintenance items from the target monitoring system's maintenance list, obtain an updated target monitoring system's maintenance list, and prioritize maintenance of the maintenance items in the risk list of the target monitoring system's maintenance items; if the risk list of the target monitoring system's maintenance items is empty, use the target monitoring system's maintenance list as the updated target monitoring system's maintenance list, and enter step S2;

[0006] S2. Update the target monitoring system to-be-maintained list to include N items P to be maintained n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained T nThe target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; the target monitoring system project priority evaluation model is used to treat the maintenance project P n Comprehensive priority assessment can maintain optimal maintenance of the target monitoring system; maintenance items are maintained based on the order in the maintenance list of the priority target monitoring system.

[0007] As a preferred technical solution of the present invention, the target monitoring system project failure risk model in step S1 includes an information extraction layer, a coordinate feature extraction layer, a risk assessment layer and a result output layer;

[0008] The information extraction layer is used to treat the maintenance project P m The coordinate information of the project to be maintained is extracted to obtain the coordinate information F m ; According to the coordinate information of the project to be maintained F m Extract the current on-site environment information and obtain the environment information H of the project to be maintained m ;

[0009] The coordinate feature extraction layer is used to maintain the project environment information H m Perform feature extraction to obtain the environmental information feature H of the project to be maintained m ';

[0010] The risk assessment layer is used to treat the environmental information characteristics of the maintenance project H m 'Perform risk assessment and obtain the risk value X of the project to be maintained m ; Set the risk value of all maintenance projects to X m The projects to be maintained that are greater than the risk threshold of the projects to be maintained are added to the risk list of the projects to be maintained in the target monitoring system;

[0011] The result output layer is used to output the risk list of projects to be maintained in the target monitoring system.

[0012] As a preferred technical solution of the present invention, the specific steps of training the risk assessment layer in the project failure risk model of the target monitoring system are as follows:

[0013] Collecting several groups of risk assessment training samples, each group of risk assessment training samples contains a group of environmental information features and corresponding risk values; combining several groups of risk assessment training samples to obtain a risk assessment training set;

[0014] The risk assessment training set is input into the target monitoring system project failure risk model to train the risk assessment layer with the risk value as the target to obtain the initial risk assessment layer; the initial risk assessment layer is model evaluated to obtain the initial risk assessment layer model evaluation result; if the initial risk assessment layer model evaluation result is passed, the initial risk assessment layer is used as the risk assessment layer in the target monitoring system project failure risk model; otherwise, the risk assessment training set is used to continue model training.

[0015] As a preferred technical solution of the present invention, the target monitoring system project priority evaluation model in step S2 includes an information feature extraction layer, a feature quantification layer, a priority evaluation layer and a result output layer;

[0016] The information feature extraction layer is used to treat the maintenance project fault information G n Perform feature extraction to obtain the fault information feature G of the item to be maintained n ', treat the maintenance project timing information T n Perform feature extraction to obtain the time series information feature T of the project to be maintained n ';

[0017] The feature quantization layer is used to calculate the fault information feature G of the item to be maintained. n 'Perform quantitative scoring to obtain the fault information score U of the item to be maintained n ; According to the time series characteristic information T of the items to be maintained n 'Perform quantitative scoring to obtain the time series information score V of the project to be maintained n ;

[0018] The priority evaluation layer is used to use the formula Q n =W1*U n +W2*V n Treat maintenance project P n Perform a priority assessment to obtain the priority score Q of the project to be maintained n ; W1 is represented by U n The weight coefficient of W2 is expressed as V n The weight coefficient of all the items to be maintained P n Arrange them from high to low according to the priority scores to obtain a list of priority target monitoring systems to be maintained;

[0019] The result output layer is used to output the priority target monitoring system to be maintained list.

[0020] As a preferred technical solution of the present invention, the specific steps of training the feature quantization layer in the project priority evaluation model of the target monitoring system include:

[0021] Collect several groups of feature quantization training samples; each group of feature quantization training individuals includes a group of fault information features and corresponding quantization scores, a group of time series information features and corresponding quantization scores; combine several groups of feature quantization training samples to obtain a feature quantization training set;

[0022] The feature quantization training set is input into the target monitoring system project priority evaluation model to train the feature quantization layer with the quantitative score as the goal to obtain the initial feature quantization layer; the initial feature quantization layer is model evaluated to obtain the initial feature quantization layer model evaluation result; if the initial feature quantization layer model evaluation result is passed, the initial feature quantization layer is used as the feature quantization layer in the target monitoring system project priority evaluation model; otherwise, the model training is continued using the feature quantization training set.

[0023] As a preferred technical solution of the present invention, the specific step of determining the specific values ​​of W1 and W2 is also included:

[0024] Construct an initial random weight coefficient iteration population and an initial elite weight coefficient iteration population; the initial random weight coefficient iteration population contains K random weight coefficient iteration individuals A k , k=1, 2, ..., K; each random weight coefficient iteration individual contains a set of randomly generated values ​​of W1 and W2; the initial elite weight coefficient iteration population contains I elite weight coefficient iteration individuals B i , i=1, 2, ..., I; each elite weight coefficient iteration individual contains a set of values ​​of W1 and W2 generated under constraints; set the maximum number of iterations E max ; Define the current number of iterations as E, E=1, 2, ..., E max ;

[0025] Collect several groups of historical maintenance project training samples, each group of historical maintenance project training samples contains the historical maintenance project sequence, historical fault information feature scores and historical time series information feature scores of all historical maintenance projects; combine several groups of historical maintenance project training samples to obtain a weight coefficient training set;

[0026] When the current number of iterations is E, the random weight coefficient in the iterative population is used to iterate individual A. k Simulate the weight coefficient training set to obtain the random fitness C k ; Iterate the elite weight coefficient of individual B in the iterative population according to the initial elite weight coefficient i The weight coefficient training set is simulated and calculated to obtain the elite fitness D i ;Use the formula Y E =sin(R*E / E max ) Get the population update dynamic ratio YE , R is the adjustment parameter; update the dynamic ratio Y according to the population E Select the elite weight coefficient iteration individuals with higher fitness in the initial elite weight coefficient iteration population and add them to the initial random weight coefficient iteration population to obtain a new initial random weight coefficient iteration population; update the dynamic ratio Y according to the population E Select random weight coefficient iteration individuals with higher fitness in the initial random weight coefficient iteration population and add them to the initial elite weight coefficient iteration population to obtain a new initial elite weight coefficient iteration population;

[0027] Reach the maximum number of iterations E max When , the optimal weight coefficient iteration individual with the largest fitness in the initial random weight coefficient iteration population and the initial elite weight coefficient iteration population is output, and the values ​​of W1 and W2 in the optimal weight coefficient iteration individual are used as the values ​​of W1 and W2 in the priority evaluation layer.

[0028] As a preferred technical solution of the present invention, the specific steps of simulating and calculating to obtain random fitness and elite fitness include:

[0029] Iterate the random weight coefficient of individual A in the initial random weight coefficient iteration population k :

[0030] The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n +W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the random weight coefficient iteration individual A k , obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the simulated historically maintained project maintenance list, calculate the accuracy of the simulated historically maintained project maintenance list and the order of the historically maintained projects, and obtain the simulation priority accuracy;

[0031] Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the random weight coefficient to iterate individual A k The random fitness C k ;

[0032] Iterate the elite weight coefficient of individual B in the initial elite weight coefficient iteration population i :

[0033] The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n+W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the elite weight coefficient iteration individual B i , obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the simulated historically maintained project maintenance list, calculate the accuracy of the simulated historically maintained project maintenance list and the order of the historically maintained projects, and obtain the simulation priority accuracy;

[0034] Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the elite weight coefficient iteration individual B i Elite fitness D i .

[0035] A highway monitoring system maintenance system, comprising:

[0036] The target system maintenance project acquisition module is used to obtain the current target monitoring system to-be-maintained list; the target monitoring system to-be-maintained list contains M to-be-maintained projects P m , m=1,2,…,M;

[0037] The risk prediction module for projects to be maintained is used to deploy the target monitoring system project failure risk model, and screen the target monitoring system to-be-maintained list based on the target monitoring system project failure risk model to obtain the target monitoring system to-be-maintained project risk list, which includes the to-be-maintained projects with monitoring risks; the target monitoring system project failure risk model is used to predict the to-be-maintained projects P m Perform risk value screening, and screen high-risk maintenance items for priority maintenance; if the risk list of the target monitoring system's maintenance items is not empty, remove the maintenance items in the risk list of the target monitoring system's maintenance items from the target monitoring system's maintenance list, obtain an updated target monitoring system's maintenance list, and prioritize maintenance of the maintenance items in the risk list of the target monitoring system's maintenance items; if the risk list of the target monitoring system's maintenance items is empty, use the target monitoring system's maintenance list as the updated target monitoring system's maintenance list, and enter step S2;

[0038] The maintenance project priority evaluation module is used to deploy the target monitoring system project priority evaluation model and update the target monitoring system maintenance list containing N maintenance projects P. n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained Tn The target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; the target monitoring system project priority evaluation model is used to treat the maintenance project P n Comprehensive priority assessment can maintain optimal maintenance of the target monitoring system; maintenance items are maintained based on the order in the maintenance list of the priority target monitoring system.

[0039] The present invention has the following advantages:

[0040] 1. The present invention can quickly identify and prioritize high-risk monitoring items through automated risk assessment and priority evaluation models, significantly improve the efficiency of maintenance work, and ensure that the monitoring system is always in the best operating state, providing strong technical support for highway traffic safety, and can adapt to different traffic conditions and environmental changes, ensuring effective monitoring system maintenance under various circumstances; through continuous monitoring and timely maintenance, traffic interruptions caused by monitoring system failures are reduced, and the stability and reliability of the monitoring system are improved.

[0041] 2. The present invention comprehensively evaluates the priority of each maintenance item by taking into account the fault information and timing information. By extracting and quantitatively scoring the features of the fault information and timing information, it can evaluate the priority of each maintenance item based on its uniqueness, which helps to quickly reorder the maintenance items and ensure that high-priority items can be responded to quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 The present invention is a schematic diagram of the structure of a highway monitoring system maintenance system used in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0044] Embodiment 1, a highway monitoring system maintenance method, comprising:

[0045] S1. Get the current target monitoring system to be maintained list; the target monitoring system to be maintained list contains M items to be maintained P m ,m=1,2,…,M;Based on the target monitoring system project failure risk model, the target monitoring system to-be-maintained list is screened to obtain the target monitoring system to-be-maintained project risk list, which includes the to-be-maintained projects with monitoring risks; The target monitoring system project failure risk model is used to select the to-be-maintained projects Pm Perform risk value screening, and screen high-risk maintenance items for priority maintenance; if the risk list of the target monitoring system's maintenance items is not empty, remove the maintenance items in the risk list of the target monitoring system's maintenance items from the target monitoring system's maintenance list, obtain an updated target monitoring system's maintenance list, and prioritize maintenance of the maintenance items in the risk list of the target monitoring system's maintenance items; if the risk list of the target monitoring system's maintenance items is empty, use the target monitoring system's maintenance list as the updated target monitoring system's maintenance list, and enter step S2;

[0046] The target monitoring system maintenance list indicates the monitoring project failures that occur simultaneously in the current period. In the highway monitoring system, there are several continuous monitoring projects. Each monitoring project monitors a certain area. The traffic volume in different areas is different, and the probability of accidents is also different.

[0047] The target monitoring system project failure risk model in step S1 includes an information extraction layer, a coordinate feature extraction layer, a risk assessment layer and a result output layer;

[0048] The information extraction layer is used to treat the maintenance project P m The coordinate information of the project to be maintained is extracted to obtain the coordinate information F m ; According to the coordinate information of the project to be maintained F m Extract the current on-site environment information and obtain the environment information H of the project to be maintained m ;

[0049] The coordinate feature extraction layer is used to maintain the project environment information H m Perform feature extraction to obtain the environmental information feature H of the project to be maintained m ';

[0050] The risk assessment layer is used to treat the environmental information characteristics of the maintenance project H m 'Perform risk assessment and obtain the risk value X of the project to be maintained m ; Set the risk value of all maintenance projects to X m The projects to be maintained that are greater than the risk threshold of the projects to be maintained are added to the risk list of the projects to be maintained in the target monitoring system; the risk threshold of the projects to be maintained is set by professional technicians based on actual conditions;

[0051] The result output layer is used to output the risk list of the target monitoring system projects to be maintained;

[0052] Specific steps for training the risk assessment layer in the project failure risk model of the target monitoring system:

[0053] Collecting several groups of risk assessment training samples, each group of risk assessment training samples contains a group of environmental information features and corresponding risk values; combining several groups of risk assessment training samples to obtain a risk assessment training set;

[0054] The risk assessment training set is input into the target monitoring system project failure risk model to train the risk assessment layer with the risk value as the target to obtain the initial risk assessment layer; the model of the initial risk assessment layer is evaluated to obtain the model evaluation result of the initial risk assessment layer; if the model evaluation result of the initial risk assessment layer is passed, the initial risk assessment layer is used as the risk assessment layer in the target monitoring system project failure risk model; otherwise, the model training is continued using the risk assessment training set;

[0055] The BP neural network model is used as the basic model of the project failure risk model of the target monitoring system. Through the information extraction layer and the coordinate feature extraction layer, the model can accurately identify the environmental characteristics of the project to be maintained, providing a detailed data basis for risk assessment; it can perform risk assessment based on the specific coordinate information and environmental information of the project to be maintained, improving the accuracy of risk prediction and environmental adaptability; it can m Sorting all maintenance items and clarifying the priority of the items will help to reasonably allocate maintenance resources. The higher the risk value, the greater the traffic volume and the more likely it is to be dangerous. When a monitoring item in a place with a higher risk value fails, it needs to be handled in a timely manner to avoid having no reliable monitoring as evidence when an accident occurs;

[0056] S2. Update the target monitoring system to-be-maintained list to include N items P to be maintained n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained T n The target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; the target monitoring system project priority evaluation model is used to treat the maintenance project P n Comprehensive priority assessment can maintain optimal maintenance of target monitoring systems; maintain maintenance items based on the order of the priority target monitoring system maintenance list;

[0057] The target monitoring system project priority evaluation model in step S2 includes an information feature extraction layer, a feature quantification layer, a priority evaluation layer and a result output layer;

[0058] The information feature extraction layer is used to treat the maintenance project fault information Gn Perform feature extraction to obtain the fault information feature G of the item to be maintained n ', treat the maintenance project timing information T n Perform feature extraction to obtain the time series information feature T of the project to be maintained n ';

[0059] The feature quantization layer is used to calculate the fault information feature G of the item to be maintained. n 'Perform quantitative scoring to obtain the fault information score U of the item to be maintained n ; According to the time series characteristic information T of the items to be maintained n 'Perform quantitative scoring to obtain the time series information score V of the project to be maintained n ;

[0060] The priority evaluation layer is used to use the formula Q n =W1*U n +W2*V n Treat maintenance project P n Perform a priority assessment to obtain the priority score Q of the project to be maintained n ; W1 is represented by U n The weight coefficient of W2 is expressed as V n The weight coefficient of all the items to be maintained P n Arrange them from high to low according to the priority scores to obtain a list of priority target monitoring systems to be maintained;

[0061] The result output layer is used to output the priority target monitoring system to be maintained list;

[0062] The specific steps of training the feature quantization layer in the project priority evaluation model of the target monitoring system include:

[0063] Collect several groups of feature quantization training samples; each group of feature quantization training individuals includes a group of fault information features and corresponding quantization scores, a group of time series information features and corresponding quantization scores; combine several groups of feature quantization training samples to obtain a feature quantization training set;

[0064] Input the feature quantization training set into the target monitoring system project priority evaluation model to train the feature quantization layer with the quantitative score as the goal to obtain the initial feature quantization layer; perform model evaluation on the initial feature quantization layer to obtain the initial feature quantization layer model evaluation result; if the initial feature quantization layer model evaluation result is passed, the initial feature quantization layer is used as the feature quantization layer in the target monitoring system project priority evaluation model; otherwise, continue model training using the feature quantization training set;

[0065] By comprehensively considering the fault information and timing information, the priority of each maintenance project is comprehensively evaluated. By extracting and quantifying the features of the fault information and timing information, the priority of each maintenance project can be evaluated based on its uniqueness, which helps to quickly reorder the maintenance projects and ensure that high-priority projects can be responded to quickly. According to the fault information, the fault causes of each maintenance project are different, such as camera failure, connection failure, etc. According to different fault causes, the maintenance time for the fault is also different. According to the actual situation, under the same road conditions, maintenance projects with short maintenance times are always given priority. During the model training process, different quantitative scores are assigned to different fault features. According to the timing information, in the same list of maintenance projects, the maintenance submissions of different maintenance projects are different. Under the same road conditions, maintenance projects that need to be maintained first are always given priority. Following the principle of timing, different quantitative scores are assigned to different timing features during the model training process to achieve the effect of comprehensive priority evaluation.

[0066] It also includes the specific steps of determining the specific values ​​of W1 and W2:

[0067] Construct an initial random weight coefficient iteration population and an initial elite weight coefficient iteration population; the initial random weight coefficient iteration population contains K random weight coefficient iteration individuals A k , k=1, 2, ..., K; each random weight coefficient iteration individual contains a set of randomly generated values ​​of W1 and W2; satisfying W1+W2=1;

[0068] The initial elite weight coefficient iteration population contains I elite weight coefficient iteration individuals B i , i=1, 2, …, I; each elite weight coefficient iteration individual contains a set of constrained generated values ​​of W1 and W2; the initial elite weight coefficient iteration population is generated within the set range, for example, W1 fluctuates in the range of 0.6-0.7, and W2 fluctuates in the range of 0.3-0.4, satisfying W1+W2=1;

[0069] By randomly generating weight coefficients, the diversity of the population is increased, which helps to explore a wider solution space. The combination of extensive exploration of the random population and fine development of the elite population balances the exploration ability and development efficiency of the algorithm. The weight coefficients of the elite population are subject to specific constraints, ensuring that the weights are set within a reasonable range and avoiding unrealistic weight distribution. The introduction of the elite population helps to maintain stability in the optimization process and maintain a certain optimization quality even when the random population exploration is not ideal. The iterative update of the weight coefficient provides the model with the ability to dynamically adjust, so that the model can adapt to the optimization needs of different stages.

[0070] Set the maximum number of iterations Emax ; Define the current number of iterations as E, E=1, 2, ..., E max ; Maximum number of iterations E max Set up by professional technicians;

[0071] Collect several groups of historical maintenance project training samples, each group of historical maintenance project training samples contains the historical maintenance project sequence, historical fault information feature scores and historical time series information feature scores of all historical maintenance projects; combine several groups of historical maintenance project training samples to obtain a weight coefficient training set;

[0072] When the current number of iterations is E, the random weight coefficient in the iterative population is used to iterate individual A. k Simulate the weight coefficient training set to obtain the random fitness C k ; Iterate the elite weight coefficient of individual B in the population according to the initial elite weight coefficient i The weight coefficient training set is simulated and calculated to obtain the elite fitness D i ;Use the formula Y E =sin(R*E / E max ) Get the population update dynamic ratio Y E , R is the adjustment parameter; update the dynamic ratio Y according to the population E Select the elite weight coefficient iteration individuals with higher fitness in the initial elite weight coefficient iteration population and add them to the initial random weight coefficient iteration population to obtain a new initial random weight coefficient iteration population; update the dynamic ratio Y according to the population E Select random weight coefficient iteration individuals with higher fitness in the initial random weight coefficient iteration population and add them to the initial elite weight coefficient iteration population to obtain a new initial elite weight coefficient iteration population;

[0073] The specific steps of simulating and calculating the random fitness and elite fitness include:

[0074] Iterate the random weight coefficient of individual A in the initial random weight coefficient iteration population k :

[0075] The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n +W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the random weight coefficient iteration individual A k, obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the simulated historically maintained project maintenance list, calculate the accuracy of the simulated historically maintained project maintenance list and the order of the historically maintained projects, and obtain the simulation priority accuracy;

[0076] Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the random weight coefficient to iterate individual A k The random fitness C k ;

[0077] Iterate the elite weight coefficient of individual B in the initial elite weight coefficient iteration population i :

[0078] The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n +W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the elite weight coefficient iteration individual B i , obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the simulated historically maintained project maintenance list, calculate the accuracy of the simulated historically maintained project maintenance list and the order of the historically maintained projects, and obtain the simulation priority accuracy;

[0079] Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the elite weight coefficient iteration individual B i Elite fitness D i ;

[0080] Reach the maximum number of iterations E max When , the optimal weight coefficient iteration individual with the largest fitness in the initial random weight coefficient iteration population and the initial elite weight coefficient iteration population is output, and the values ​​of W1 and W2 in the optimal weight coefficient iteration individual are used as the values ​​of W1 and W2 in the priority evaluation layer;

[0081] Through iterative population construction and simulation calculation, the weight distribution of fault information and timing information is optimized, which more accurately reflects the importance of each feature in priority evaluation; the simulation calculation using historical maintenance project data improves the accuracy of priority scoring, making the order of maintenance work more in line with the actual situation; the population updates the dynamic proportion Y EThe introduction of enables the iterative population to be dynamically adjusted according to the progress of the current iteration, which enhances the adaptability of the model, expands the range of the optimal solution, and makes it easier to quickly approach the optimal solution. By setting up two different iterative populations, it not only ensures the population diversity in the early stage of iteration and diversifies the solutions of the weight coefficients, but also makes it converge quickly to the optimal solution in the process of increasing the number of iterations, thereby improving the search efficiency of the optimal solution.

[0082] Example 2, a highway monitoring system maintenance system, see Figure 1 As shown, including:

[0083] The target system maintenance project acquisition module is used to obtain the current target monitoring system to-be-maintained list; the target monitoring system to-be-maintained list contains M to-be-maintained projects P m , m=1,2,…,M;

[0084] The risk prediction module for projects to be maintained is used to deploy the target monitoring system project failure risk model, and screen the target monitoring system to-be-maintained list based on the target monitoring system project failure risk model to obtain the target monitoring system to-be-maintained project risk list, which includes the to-be-maintained projects with monitoring risks; the target monitoring system project failure risk model is used to predict the to-be-maintained projects P m Perform risk value screening, and screen high-risk maintenance items for priority maintenance; if the risk list of the target monitoring system's maintenance items is not empty, remove the maintenance items in the risk list of the target monitoring system's maintenance items from the target monitoring system's maintenance list, obtain an updated target monitoring system's maintenance list, and prioritize maintenance of the maintenance items in the risk list of the target monitoring system's maintenance items; if the risk list of the target monitoring system's maintenance items is empty, use the target monitoring system's maintenance list as the updated target monitoring system's maintenance list, and enter step S2;

[0085] The maintenance project priority evaluation module is used to deploy the target monitoring system project priority evaluation model and update the target monitoring system maintenance list containing N maintenance projects P. n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained T n The target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; the target monitoring system project priority evaluation model is used to treat the maintenance project P nComprehensive priority assessment can maintain optimal maintenance of the target monitoring system; maintenance items are maintained based on the order in the maintenance list of the priority target monitoring system.

[0086] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A highway monitoring system maintenance method, characterized in that: include: S1. Obtain the current target monitoring system to be maintained list; The target monitoring system maintenance list contains M items to be maintained P m , m=1,2,…,M; Based on the target monitoring system project failure risk model, the target monitoring system to-be-maintained list is screened to obtain a target monitoring system to-be-maintained project risk list, wherein the target monitoring system to-be-maintained project risk list includes to-be-maintained projects with monitoring risk conditions; The project failure risk model of the target monitoring system is used to treat the maintenance project P m Perform risk value screening and select high-risk maintenance items for priority maintenance; If the target monitoring system to-be-maintained project risk list is not empty, the to-be-maintained projects in the target monitoring system to-be-maintained project risk list are removed from the target monitoring system to-be-maintained project list to obtain an updated target monitoring system to-be-maintained project list, and the to-be-maintained projects in the target monitoring system to-be-maintained project risk list are maintained preferentially; If the target monitoring system to-be-maintained project risk list is empty, the target monitoring system to-be-maintained list is used as the updated target monitoring system to-be-maintained list, and the process goes to step S2; S2. Update the target monitoring system to-be-maintained list to include N items P to be maintained n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained T n The target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; The project priority evaluation model of the target monitoring system is used to treat the maintenance project P n Comprehensive priority assessment can maintain optimal maintenance of target monitoring systems; maintain maintenance items based on the order of the priority target monitoring system maintenance list; The target monitoring system project failure risk model in step S1 includes an information extraction layer, a coordinate feature extraction layer, a risk assessment layer and a result output layer; The information extraction layer is used to treat the maintenance project P m Extract the coordinate information of the project to be maintained to obtain the coordinate information F m ; According to the coordinate information of the project to be maintained F m Extract the current on-site environment information and obtain the environment information H of the project to be maintained m ; The coordinate feature extraction layer is used to maintain the project environment information H m Perform feature extraction to obtain the environmental information feature H of the project to be maintained m '; The risk assessment layer is used to treat the environmental information characteristics of the maintenance project H m 'Perform risk assessment and obtain the risk value X of the project to be maintained m ; Set the risk value of all maintenance projects to X m The projects to be maintained that are greater than the risk threshold of the projects to be maintained are added to the risk list of the projects to be maintained in the target monitoring system; The result output layer is used to output the risk list of the target monitoring system projects to be maintained; The target monitoring system project priority evaluation model in step S2 includes an information feature extraction layer, a feature quantification layer, a priority evaluation layer and a result output layer; The information feature extraction layer is used to treat the maintenance project fault information G n Perform feature extraction to obtain the fault information feature G of the item to be maintained n ', treat the maintenance project timing information T n Perform feature extraction to obtain the time series information feature T of the project to be maintained n '; The feature quantization layer is used to calculate the fault information feature G of the item to be maintained n 'Perform quantitative scoring to obtain the fault information score U of the item to be maintained n ; According to the time series characteristic information T of the items to be maintained n 'Perform quantitative scoring to obtain the time series information score V of the project to be maintained n ; The priority evaluation layer is used to use the formula Q n =W1*U n +W2*V n Treat maintenance project P n Perform a priority assessment to obtain the priority score Q of the project to be maintained n ; W1 is represented by U n The weight coefficient of W2 is expressed as V n The weight coefficient of all the items to be maintained P n Arrange them from high to low according to the priority scores to obtain a list of priority target monitoring systems to be maintained; The result output layer is used to output the priority target monitoring system to be maintained list; It also includes the specific steps of determining the specific values ​​of W1 and W2: Construct an initial random weight coefficient iteration population and an initial elite weight coefficient iteration population; the initial random weight coefficient iteration population contains K random weight coefficient iteration individuals A k , k = 1, 2, ..., K; each random weight coefficient iteration individual contains a set of randomly generated values ​​of W1 and W2; the initial elite weight coefficient iteration population contains I elite weight coefficient iteration individuals B i , i = 1, 2, ..., I; each elite weight coefficient iteration individual contains a set of constrained generated values ​​of W1 and W2; set the maximum number of iterations E max ; Define the current number of iterations as E, E = 1, 2, ..., E max ; Collect several groups of historical maintenance project training samples, each group of historical maintenance project training samples contains the historical maintenance project sequence, historical fault information feature scores and historical time series information feature scores of all historical maintenance projects; Combine several groups of historically maintained project training samples to obtain a weight coefficient training set; When the current number of iterations is E, the random weight coefficient in the iterative population is used to iterate individual A. k Simulate the weight coefficient training set to obtain the random fitness C k ; Iterate the elite weight coefficient of individual B in the iterative population according to the initial elite weight coefficient i The weight coefficient training set is simulated and calculated to obtain the elite fitness D i ;Use the formula Y E =sin(R*E / E max ) Get the population update dynamic ratio Y E , R is the adjustment parameter; update the dynamic ratio Y according to the population E Select the elite weight coefficient iteration individuals with higher fitness in the initial elite weight coefficient iteration population and add them to the initial random weight coefficient iteration population to obtain a new initial random weight coefficient iteration population; update the dynamic ratio Y according to the population E Select random weight coefficient iteration individuals with higher fitness in the initial random weight coefficient iteration population and add them to the initial elite weight coefficient iteration population to obtain a new initial elite weight coefficient iteration population; Reach the maximum number of iterations E max When , the optimal weight coefficient iteration individual with the largest fitness in the initial random weight coefficient iteration population and the initial elite weight coefficient iteration population is output, and the values ​​of W1 and W2 in the optimal weight coefficient iteration individual are used as the values ​​of W1 and W2 in the priority evaluation layer.

2. A highway monitoring system maintenance method according to claim 1, characterized in that: Specific steps for training the risk assessment layer in the project failure risk model of the target monitoring system: Collect several groups of risk assessment training samples, each group of risk assessment training samples contains a group of environmental information features and corresponding risk values; Combining several groups of risk assessment training samples to obtain a risk assessment training set; The risk assessment training set is input into the target monitoring system project failure risk model to train the risk assessment layer with the risk value as the target, and an initial risk assessment layer is obtained; Performing model evaluation on the initial risk assessment layer to obtain the model evaluation result of the initial risk assessment layer; If the evaluation result of the initial risk assessment layer model is passed, the initial risk assessment layer will be used as the risk assessment layer in the target monitoring system project failure risk model; Otherwise, continue model training using the risk assessment training set.

3. A highway monitoring system maintenance method according to claim 2, characterized in that: The specific steps of training the feature quantization layer in the project priority evaluation model of the target monitoring system include: Collect several groups of feature quantization training samples; each group of feature quantization training individuals includes a group of fault information features and corresponding quantization scores, a group of time series information features and corresponding quantization scores; combine several groups of feature quantization training samples to obtain a feature quantization training set; The feature quantization training set is input into the target monitoring system project priority evaluation model to train the feature quantization layer with the quantitative score as the goal to obtain the initial feature quantization layer; the initial feature quantization layer is model evaluated to obtain the initial feature quantization layer model evaluation result; if the initial feature quantization layer model evaluation result is passed, the initial feature quantization layer is used as the feature quantization layer in the target monitoring system project priority evaluation model; otherwise, the model training is continued using the feature quantization training set.

4. A highway monitoring system maintenance method according to claim 3, characterized in that: The specific steps of simulating and calculating the random fitness and elite fitness include: Iterate the random weight coefficient of individual A in the initial random weight coefficient iteration population k : The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n +W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the random weight coefficient iteration individual A k , obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the maintenance list of the simulated historically maintained projects, calculate the accuracy of the maintenance list of the simulated historically maintained projects and the order of the historically maintained projects, and obtain the simulation priority accuracy; Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the random weight coefficient to iterate individual A k The random fitness C k ; Iterate the elite weight coefficient of individual B in the initial elite weight coefficient iteration population i : The weight coefficient of the historical maintenance project training set is used for all historical maintenance projects in the training sample. n =W1*U n +W2*V n For simulation calculation, the values ​​of W1 and W2 are derived from the elite weight coefficient iteration individual B i , obtain the simulation priority scores corresponding to the historically maintained projects, sort all the historically maintained projects according to the simulation priority scores corresponding to them, obtain the maintenance list of the simulated historically maintained projects, calculate the accuracy of the maintenance list of the simulated historically maintained projects and the order of the historically maintained projects, and obtain the simulation priority accuracy; Calculate the average of the simulated priority accuracy of several groups of historically maintained project training samples in the weight coefficient training set, and use the average of the simulated priority accuracy as the elite weight coefficient iteration individual B i Elite fitness D i .

5. A highway monitoring system maintenance system, characterized in that: The system applies a highway monitoring system maintenance method according to any one of claims 1 to 4, including: The target system maintenance project acquisition module is used to obtain the current target monitoring system to-be-maintained list; the target monitoring system to-be-maintained list contains M to-be-maintained projects P m , m=1,2,…,M; The risk prediction module for projects to be maintained is used to deploy the target monitoring system project failure risk model, and screen the target monitoring system to-be-maintained list based on the target monitoring system project failure risk model to obtain the target monitoring system to-be-maintained project risk list, which includes the to-be-maintained projects with monitoring risks; the target monitoring system project failure risk model is used to predict the to-be-maintained projects P m Perform risk value screening, and screen high-risk maintenance items for priority maintenance; if the risk list of the target monitoring system's maintenance items is not empty, remove the maintenance items in the risk list of the target monitoring system's maintenance items from the target monitoring system's maintenance list, obtain an updated target monitoring system's maintenance list, and prioritize maintenance of the maintenance items in the risk list of the target monitoring system's maintenance items; if the risk list of the target monitoring system's maintenance items is empty, use the target monitoring system's maintenance list as the updated target monitoring system's maintenance list, and enter step S2; The maintenance project priority evaluation module is used to deploy the target monitoring system project priority evaluation model and update the target monitoring system maintenance list containing N maintenance projects P. n , obtain the fault information of the item to be maintained G n And the time sequence information of the items to be maintained T n , based on all the projects to be maintained P n Fault information of items to be maintained G n , Time sequence information of items to be maintained T n The target monitoring system project priority evaluation model is used to reorder the target monitoring system maintenance list to obtain a priority target monitoring system maintenance list; the target monitoring system project priority evaluation model is used to treat the maintenance project P n Comprehensive priority assessment can maintain optimal maintenance of the target monitoring system; maintenance items are maintained based on the order in the maintenance list of the priority target monitoring system.

Images