Unmanned aerial vehicle cooperative inspection path planning and airport layout method and system for highway multi-facility maintenance

Abstract

The application discloses a kind of unmanned aerial vehicle cooperative inspection path planning and airport layout method and system for highway multi-facility maintenance.The method constructs multidimensional environment representation system by collecting highway multi-source data, establishes comprehensive optimization model with coverage range, response time and total cost as target, generates airport layout scheme using adjustable service radius strategy and heuristic search algorithm, and extracts optimal equilibrium solution set through multi-objective sorting screening.Based on facility type and maintenance urgency, task priority is determined, and multi-unmanned aerial vehicle cooperative path planning model is established to plan conflict-free cyclic inspection path.Predictive control strategy is used for online rolling replanning, and system closed-loop adaptive optimization is realized by integrating inspection data feedback.The application realizes the cooperative optimization of airport layout and inspection path, improves the inspection efficiency, reduces the operating cost, and enhances the adaptive ability of the system to respond to dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) application technology, and in particular to a method and system for collaborative inspection path planning and airport deployment of UAVs for highway multi-facility maintenance. Background Technology

[0002] As my country's expressway network continues to expand, the demand for daily inspection and maintenance of its roadside facilities (such as roads, bridges, slopes, and photovoltaic panels) is increasing. Traditional manual inspection methods suffer from low efficiency, high cost, and significant safety risks, especially in areas with complex terrain and scattered distribution, where effective coverage is difficult. Drone technology, with its advantages of high flexibility, strong accessibility, and relatively low cost, has become an important tool for infrastructure inspection.

[0003] However, applying drones to highway multi-facility maintenance scenarios still faces numerous challenges. First, the highway environment is complex, and inspection tasks are subject to multiple constraints such as geographical terrain, airspace control, and weather conditions, making traditional single-path planning methods unsuitable. Second, the layout of drone airports directly affects inspection efficiency and operating costs. Existing research often uses static coverage models for site selection, neglecting the dynamic coupling relationship between drone endurance, service radius, and path planning, making it difficult to achieve a balanced optimization between coverage, response time, and construction costs. Third, in the face of sudden weather events and temporary airspace restrictions, there is a lack of effective online replanning mechanisms, failing to guarantee the continuity and safety of inspection tasks. Furthermore, most existing solutions lack self-learning and continuous optimization capabilities based on historical inspection data, and the overall system efficiency needs further improvement.

[0004] Therefore, there is an urgent need in this field for a highway unmanned aerial vehicle (UAV) intelligent inspection solution that can deeply integrate environmental constraints, achieve integrated and collaborative optimization of airport deployment and inspection paths, and possess dynamic response and continuous learning capabilities. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and provide a method and system for UAV collaborative inspection path planning and airport deployment for highway multi-facility maintenance, so as to achieve collaborative optimization of airport layout and inspection path, improve inspection efficiency and resource utilization, and enhance the system's adaptability and robustness in dynamic environments.

[0006] In a first aspect, embodiments of this application provide a method for UAV collaborative inspection path planning and airport deployment for multi-facility maintenance of highways, the method comprising:

[0007] S1: Collect data on highway network, facility distribution, topography, airspace restrictions, and meteorology to construct a multi-dimensional environmental characterization system;

[0008] S2: Establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. Use an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes. Extract the optimal equilibrium solution set through multi-objective sorting and filtering, and determine the final airport spatial configuration scheme accordingly.

[0009] S3: Determine the priority sequence for inspection tasks based on facility type, history of damage, and urgency of maintenance;

[0010] S4: Establish a multi-UAV collaborative trajectory planning model, assign inspection tasks to each UAV, and plan a cyclical inspection path that meets the endurance requirements and is free from conflicts.

[0011] S5: Based on real-time environmental changes, a predictive control optimization strategy is used to perform online rolling updates and trajectory adjustments on the collaborative trajectory planning model;

[0012] S6: Integrate inspection data to generate facility status reports, and feed back historical performance data to the models in steps S2 and S4 to achieve adaptive adjustment and continuous optimization of system parameters.

[0013] Secondly, embodiments of this application provide a UAV collaborative inspection path planning and airport deployment system for highway multi-facility maintenance, applied to the UAV collaborative inspection path planning and airport deployment method for highway multi-facility maintenance as described in the first aspect. The system includes:

[0014] The environmental characterization system construction module is used to collect data on highway network, facility distribution, topography, airspace restrictions, and meteorology to construct a multi-dimensional environmental characterization system.

[0015] The airport layout optimization module is used to establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. It uses an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes, and extracts the optimal equilibrium solution set through multi-objective sorting and filtering to determine the final airport spatial configuration scheme.

[0016] The task sequence planning module is used to determine the execution priority sequence of inspection tasks based on facility type, history of damage, and maintenance urgency.

[0017] The collaborative path planning module is used to establish a multi-UAV collaborative trajectory planning model, assign inspection tasks to each UAV, and plan a cyclical inspection path that meets the endurance requirements and is free from conflicts.

[0018] The dynamic replanning module is used to perform online rolling updates and trajectory adjustments to the collaborative trajectory planning model based on real-time environmental changes and using predictive control optimization strategies.

[0019] The closed-loop optimization module is used to integrate inspection data to generate facility status reports and feed historical performance data back to the models in steps S2 and S4 to achieve adaptive adjustment and continuous optimization of system parameters.

[0020] Thirdly, embodiments of this application provide an electronic device, including:

[0021] processor;

[0022] Memory used to store processor-executable instructions;

[0023] The processor is configured to implement, when executing the instructions, the method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for highway multi-facility maintenance as described in the first aspect.

[0024] Fourthly, embodiments of this application provide a computer-readable storage medium storing a program that instructs a device to execute the UAV collaborative inspection path planning and airport deployment method for highway multi-facility maintenance as described in the first aspect.

[0025] The beneficial effects of this invention are as follows:

[0026] 1. Improved inspection efficiency and resource utilization: By establishing a multi-objective comprehensive optimization model for coverage, response time and total cost, and by adopting an adjustable service radius strategy and heuristic search algorithm, the scientific planning of airport spatial layout was achieved, significantly expanding the coverage of a single inspection, shortening emergency response time, optimizing the allocation of drones and airport resources, and overcoming the problems of uneven coverage and slow response caused by traditional experience-based deployment.

[0027] 2. Enhanced system intelligence and adaptability: By constructing a multi-dimensional environmental characterization system that integrates geospatial, airspace, meteorological, and facility distribution, and by adopting an online rolling optimization and replanning strategy based on predictive control, the system can dynamically perceive environmental changes (such as weather and airspace restrictions) and adjust the UAV trajectory in real time, effectively avoiding risks and ensuring the safe and smooth execution of inspection tasks in complex environments.

[0028] 3. Achieved closed-loop optimization and continuous improvement: By integrating multi-source inspection data and establishing a historical performance data feedback mechanism, the system can dynamically assess the health status of facilities and the operational efficiency of the system, and automatically adjust and optimize model weight parameters and path planning constraints, forming a complete closed loop of "planning-execution-evaluation-feedback-optimization". This enables the system to learn itself and continuously improve its performance throughout its entire life cycle, significantly reducing long-term operation and maintenance costs. Attached Figure Description

[0029] Figure 1This is a schematic diagram of a method for collaborative inspection path planning and airport deployment for highway multi-facility maintenance provided in an embodiment of this application.

[0030] Figure 2 The system architecture diagram for UAV collaborative inspection path planning and airport deployment for highway multi-facility maintenance provided in this application.

[0031] Figure 3 A schematic diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0033] It should be noted that in the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0034] Based on the embodiments described in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] Example 1

[0036] Figure 1 This is a schematic diagram illustrating a method for collaborative inspection path planning and airport deployment using unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways, provided as an embodiment of this application. Figure 1 As shown, a method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways includes:

[0037] S1. Collect highway network, facility distribution, topography, airspace restrictions, and meteorological data to establish an environmental characterization system encompassing geospatial constraints, airspace control constraints, meteorological dynamic constraints, and facility distribution constraints. Collect multi-source data related to highways (road network, facilities, topography, airspace, meteorology) and construct an environmental characterization system including various constraints, such as geospatial constraints (topography, slope, obstacles); airspace control constraints (no-fly zones, restricted flight zones); meteorological dynamic constraints (wind speed, precipitation, visibility); and facility distribution constraints (road surface, bridges, slopes, photovoltaic areas, etc.). This provides dynamic and accurate environmental foundation data for subsequent airport deployment and route planning.

[0038] Specifically, in this embodiment, highway network topology, facility spatial distribution, digital elevation, airspace control, and multidimensional meteorological data are collected through a GIS platform, IoT sensors, and meteorological monitoring stations. The data acquisition function clearly defines the source channels and data types of the data required to construct the environmental characterization system. Sources: GIS platform (geographic information), IoT sensors (real-time facility status), meteorological monitoring stations (real-time weather). Data types: highway network topology, facility spatial distribution, digital elevation (topography), airspace control information, and multidimensional meteorological data. This ensures the comprehensiveness, accuracy, and real-time nature of the basic data, providing a reliable data foundation for subsequent analysis.

[0039] Based on collected multi-source data, a multi-dimensional environmental characterization system is constructed, incorporating the following constraints: Geospatial constraints: including terrain undulation, slope and aspect, and obstacle distribution; Airspace control constraints: including no-fly zones, restricted-fly zones, and route reservation areas; Meteorological dynamic constraints: including real-time wind speed, precipitation, and visibility; Facility distribution constraints: including the spatial distribution density of road sections, bridges, slopes, and photovoltaic areas. The multi-dimensional constraint system construction function transforms and categorizes the collected multi-source data into a series of specific constraints that can be directly processed by mathematical models. Specifically, it defines four categories of constraints: Geospatial constraints: transforming terrain data into quantifiable flight obstacle indicators such as terrain undulation and slope and aspect; Airspace control constraints: transforming airspace data into explicit rule-based constraints such as no-fly zones, restricted-fly zones, and route reservation areas; Meteorological dynamic constraints: transforming meteorological data into key dynamic parameters affecting UAV flight, such as real-time wind speed, precipitation, and visibility; and Facility distribution constraints: transforming facility location data into spatial distribution density information for assessing coverage requirements. By structuring and standardizing the raw data, a "rule base" is formed that computer algorithms can understand and process, serving as a bridge connecting the real world and digital models.

[0040] This system utilizes data fusion technology to perform spatiotemporal registration and consistency processing on multi-source heterogeneous environmental data, establishing a mapping relationship between environmental data and UAV performance parameters. This provides dynamically updated environmental foundation data for airport spatial configuration optimization and flight path planning. The data fusion and processing function integrates and standardizes environmental data from different sources and in different formats (heterogeneous), and correlates it with UAV performance. Spatiotemporal registration ensures that all data are aligned and consistent in timestamps and geographic coordinates. Consistency processing resolves potential conflicts or errors in multi-source data, forming a unified and reliable dataset. Establishing a mapping relationship links environmental data (such as wind speed) with UAV performance parameters (such as maximum wind resistance level and endurance). For example, strong winds reduce the effective flight speed of UAVs. .

[0041] S2. Establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. Employ an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes. Extract the optimal equilibrium solution set through multi-objective ranking and filtering. The airport layout optimization model is constructed and candidate schemes are generated by establishing a multi-objective optimization model with coverage, response time, and total cost as optimization objectives. The optimal location, number, and service radius of airports are determined to achieve the optimal balance between economy and efficiency.

[0042] The multi-objective optimization function construction function establishes a formal mathematical model, quantifying the three high-level objectives of "maximizing coverage, minimizing response time, and minimizing total cost" into computable mathematical functions. This provides clear and measurable optimization goals and directions for the entire optimization process. This is a crucial step from qualitative requirements to quantitative calculations.

[0043] Specifically, in this embodiment, a multi-objective optimization function is constructed to simultaneously optimize three objectives: coverage, response time, and total cost. Three quantifiable objectives are defined: maximizing coverage, minimizing response time, and minimizing total cost, transforming the abstract requirements into a computable mathematical problem.

[0044] The construction of the multi-objective optimization function includes:

[0045] Establish the objective function for maximizing comprehensive coverage effectiveness, expressed as:

[0046] ,

[0047] in, This represents the overall coverage effectiveness. A larger value for this objective function indicates better overall coverage. This indicates the total number of facilities. It is used to calculate average coverage effectiveness, normalize the results, and facilitate comparisons across different scenarios. Facilities The importance weighting is such that important facilities (such as large bridges) have a higher weight than minor facilities (such as ordinary road sections), and they account for a larger proportion in the coverage assessment. It is the coverage sensitivity coefficient, a constant greater than 0, used to adjust coverage quality. Sensitivity to the final performance contribution. The larger the coverage area, the more significant the performance gain from even a small improvement in coverage quality. Indicates drone facilities Coverage quality, a value between 0 and 1, indicates the quality of coverage, not just whether or not coverage occurs. It is a drone facilities The effective distance between them. This usually refers to the actual flight distance after taking into account terrain undulations, rather than the straight-line distance. It is the distance attenuation coefficient. As an indicator function, when a certain drone exists Make It takes a value of 1 when the facility is covered and 0 otherwise. It is a conditional function that ensures that only covered facilities are included in the performance calculation.

[0048] The overall coverage effectiveness maximization objective function is used to measure and maximize the overall coverage effect of the drone airport network on highway-side facilities. Differentiated coverage: This is achieved by introducing facility importance weights. This ensures that critical infrastructure (such as large bridges and steep slopes) receives higher coverage priority and better coverage quality than ordinary road sections. This avoids a "one-size-fits-all" coverage strategy. Coverage quality assessment: using functions. To assess the "quality" of coverage, not just "present or absent." Facilities closer to the airport generally have better coverage quality. The closer to 1 (perfect coverage), the better; however, coverage quality decreases exponentially with increasing distance. This aligns better with engineering practices. Effective coverage is guaranteed by an indicator function. Set a minimum coverage quality threshold A facility is considered "effectively covered" and included in performance calculations only if it can be covered by at least one airport with a quality exceeding this threshold. This ensures the practicality of coverage. Normalization: Divide by the total number of facilities. This makes the result an average performance value, which is convenient for comparison of road networks of different sizes.

[0049] The objective function for spatiotemporal response performance is established as follows:

[0050] ,

[0051] in, This represents the spatiotemporal response performance. The smaller the value of this objective function, the shorter the average response time and the stronger the response capability. This represents the effective flight speed of the drone. It is the average speed after considering the drone's own cruising speed and basic performance. For path The meteorological impact factor is a quantitative value that reflects the degree to which current meteorological conditions (such as headwinds and rainfall) affect the drone's flight speed along its path. The larger the value, the greater the negative impact. Meteorological sensitivity coefficient, and satisfying This ensures that the denominator is positive, thus maintaining the mathematical validity of the formula; This indicates that all drones The minimum value operation. This means that the response time of a facility is determined by its fastest responder (i.e., the drone that can arrive the fastest).

[0052] The spatiotemporal response performance minimization objective function measures and minimizes the weighted average response time of a UAV traveling from an airport to any facility point under the worst-case scenario. The fastest response principle: the operator `mini` means that the response time of a facility is determined by the nearest and fastest UAV airport. This naturally drives the deployment of more airports in densely populated areas. Environmental impact modeling: [Item] This innovative approach quantifies and incorporates the negative impact of real-time weather conditions (such as headwinds and heavy rain) on flight speed into the model. The worse the weather (…), the higher the flight speed. The larger the value, the lower the effective speed and the longer the expected response time. Prioritize critical facilities: also introduce weights. This allows reducing response time to critical facilities to play a larger role in the optimization objectives. Normalization: Similarly, by dividing by... The average response efficiency is obtained.

[0053] Establish the life-cycle cost objective function, expressed as:

[0054] ,

[0055] in, This represents the total lifecycle cost. A smaller value for this objective function indicates a lower total project cost. For the airport Construction costs (one-time initial investment). Its annual maintenance cost, For drones Purchase cost (one-time initial investment). For drones Annual energy consumption (e.g., electricity consumption). For the first The annual energy price is used to convert energy consumption into costs. It is the discount rate, used to discount future costs to their present value, reflecting the time value of money. Project duration; the total planned project lifecycle. The variable is the time variable (year). The summation expression iterates from 1 to... .

[0056] The objective function for minimizing the entire lifecycle cost is used to measure and minimize the total economic cost of the entire UAV inspection system from construction to operation and maintenance. Comprehensive cost accounting: This considers not only one-time investments (airport construction costs)... The cost of purchasing drones It also takes into account long-term operating costs (airport annual maintenance costs). Annual energy consumption cost of drones The time value of money: through the discount rate. This discounts future annual operating and maintenance costs to their present value. This makes economic assessments more scientific and accurate, avoiding underestimation of long-term operating costs. Project lifecycle management: explicitly considers the total cost over the entire project lifecycle T, supporting long-term investment decisions.

[0057] A decomposition-based multi-objective evolutionary algorithm is employed. This algorithm decomposes the multi-objective optimization problem into multiple single-objective sub-problems and uses a dynamic resource allocation strategy to coordinate the search processes among these sub-problems. Ultimately, a uniformly distributed optimal solution set is obtained, and collaborative optimization is performed using the following comprehensive utility function:

[0058] ,

[0059] in, This represents the overall utility value. It is used in decompositional multi-objective evolutionary algorithms to coordinate the search of different subproblems and guide the algorithm towards a direction with better overall performance. These are weighting coefficients, representing the relative importance of coverage, response, and cost objectives in the overall utility. They are set by the decision-maker and must satisfy... , These are the reference maximum values ​​for each objective function. They are used to normalize three objective functions that may have completely different dimensions and orders of magnitude, allowing them to be added together for comparison. These values ​​are typically determined through pre-estimation or historical maximum values ​​during the iteration process.

[0060] Final comprehensive utility function Its function is to provide a coordination mechanism for these three competing objectives. This is achieved by adjusting the weighting coefficients. Decision-makers can express their preferences (e.g., whether they value coverage more or cost control more), thereby guiding optimization algorithms to find the optimal trade-off solution that meets specific needs.

[0061] An adjustable service radius strategy is adopted, dynamically adjusting the service radius value within a preset service radius range, and iteratively generating candidate airport layout schemes using a heuristic set coverage search algorithm. This step provides an efficient search method to intelligently generate a set of candidate airport layout schemes with different trade-offs between coverage, response time, and cost from a vast solution space, providing rich and high-quality input for subsequent multi-objective optimization screening.

[0062] The adjustable service radius strategy dynamically adjusts the service radius value within a preset service radius range, and iteratively generates airport layout candidate schemes using a heuristic set coverage search algorithm, including:

[0063] Set the search range for the service radius: ,in To minimize the service radius, To maximize the service radius, based on the drone's endurance Effective flight speed and the facility's maximum permissible response time The calculation determines the radius value; ensuring it conforms to actual physical limitations. The adjustable service radius strategy is implemented for dynamic parameter exploration. The search range for the service radius is set. , ], and clearly define its composition based on the physical performance of the drone (endurance) ,speed ) and business requirements (maximum response time) The calculations determine that this strategy allows the algorithm to explore different coverage strategies. Small radius: tends to generate a large number of high-density airport layout schemes (high coverage quality, fast response, but high cost). Large radius: tends to generate a small number of low-density airport layout schemes (low cost, but coverage quality and response speed may decrease). This ensures the diversity of generated candidate schemes, covering various possible scenarios from "high cost-high efficiency" to "low cost-low efficiency".

[0064] For each candidate service radius , refers to the geographic radius from which a drone can effectively provide service (coverage facilities) from a drone airport. This is the core variable that is dynamically adjusted in the algorithm. The following iterative process is executed:

[0065] a) Initialize the set of uncovered target points U (a dynamically updated set that initially contains all facilities to be covered. The algorithm continuously removes covered facilities from U, and terminates when U is empty). This set contains all facilities to be covered, with each facility having a weight. .

[0066] b) When U is not empty, select the candidate airport location that can cover the most uncovered target points. :

[0067] ,

[0068] in, For position The set of unsupported facilities that can be covered, i.e., located in the location Facilities located within a circle of radius R that are not yet covered by other airports. To improve coverage quality, a more refined evaluation of coverage effectiveness is needed, going beyond simply whether or not coverage is achieved in binary form. In each iteration, the algorithm selects the location that covers the weighted sum of the most uncovered target points as the deployment point for the new airport. B is the set of all candidate airport locations. This typically consists of a pre-selected series of potential airport deployment points along a highway. For position facilities The distance between them.

[0069] c) Add to the airport set and remove all that meet the criteria from U. facilities This algorithm efficiently solves the NP-hard set covering problem, ensuring the generation of high-quality candidate solutions within a reasonable time.

[0070] After each airport location selection, the overall coverage effectiveness of the current solution is calculated based on the objective function of maximizing comprehensive coverage effectiveness, the objective function of spatiotemporal response effectiveness, and the objective function of total lifecycle cost. (Calculated after each airport selection to evaluate the quality of the current partial solution), Spatiotemporal response performance (Calculated after each airport selection) and total lifecycle cost (Calculated after each airport is selected. Cost increases with the number of airports); After traversing all candidate service radii, output a set of candidate solutions. This allows for subsequent multi-objective optimization. It provides a computationally efficient way to quickly find a near-optimal (though not absolutely optimal) feasible airport deployment scheme for any given service radius R. This is a practical method for solving NP-hard problems.

[0071] The candidate solutions are optimized through multi-objective sorting and screening, the optimal equilibrium solution set is extracted, and the corresponding airport location, number and service radius configuration are recorded for each optimal equilibrium solution.

[0072] Heuristic algorithms generate a large number of high-quality solutions within a reasonable timeframe, avoiding the computational explosion problem of exhaustive search. An adjustable service radius strategy ensures that the generated candidate solutions have a wide performance distribution (coverage, response, cost), covering a variety of possible trade-offs. Each solution is equipped with precise quantitative metrics, making comparisons and decisions between solutions scientific and objective. Its output (the candidate solution set) seamlessly connects the front-end mathematical model with the subsequent multi-objective ranking and selection process, serving as a crucial link in the entire optimization workflow.

[0073] The process involves multi-objective ranking and screening to optimize candidate schemes, extracting the optimal equilibrium solution set, and immediately calculating three key performance indicators for the current (partial) schemes after selecting a new airport in each iteration. Specifically, it includes the following steps:

[0074] Establish an evaluation vector for candidate solutions: For each candidate solution (Representing a specific airport deployment configuration (such as airport location, number, and service radius)), an evaluation vector is constructed based on the objective function:

[0075] ;

[0076] in, For the plan The evaluation vector is a multi-dimensional vector, where each component corresponds to the value of an optimization objective, used to comprehensively and quantitatively evaluate the merits of a solution. For the plan The overall coverage effectiveness achieved. plan The achieved spatiotemporal response performance (usually related to response time, the smaller the value, the better). For the plan The algorithm calculates the total lifecycle cost (the smaller the value, the better). It tightly integrates solution generation with performance evaluation. While constructing solutions, the algorithm can understand the performance of each solution in real time, providing a direct data foundation for subsequent ranking and selection. It condenses the complex performance of each solution into a standardized, comparable three-dimensional data point, laying the foundation for subsequent mathematical comparison and selection. It transforms solutions from "engineering descriptions" into "mathematical objects."

[0077] Define the dominance relationship between alternatives: for any two candidate alternatives and If the following conditions are met simultaneously:

[0078] ,

[0079] ,

[0080] ,

[0081] If at least one inequality is strictly true, then the scheme is called a scheme. Domination Plan That is, if the plan Not inferior to in any goal (That is, better or equal coverage, faster or equal response, lower or equal cost), and strictly better in at least one objective, then it is called... Dominate A solution that is not dominated by any other solution is called the optimal solution. By comparing each solution pairwise, all solutions not dominated by any other solution are selected, forming an efficient solution set. This is the core selection step. It ensures that every solution remaining is an indisputable leader, with no absolute superiority or inferiority among them. For example, solution A (good coverage, high cost) and solution B (poor coverage, low cost) may coexist in this set because they are not mutually dominant, representing different trade-off strategies.

[0082] Selecting an efficient solution set: Identify all solutions from all candidate solutions that are not dominated by other solutions, forming an efficient solution set. (Efficient solution set. This is the set of all non-dominated solutions selected from all candidate solutions. This set represents all possible optimal trade-offs among the three objectives of coverage, response, and cost.) After traversing all candidate service radii, the algorithm outputs a set of candidate solutions.

[0083] Calculate the solution set distribution index: for efficient solution sets Calculate its hypervolume index HV and spacing index SP:

[0084] ,

[0085] ,

[0086] in, , , These are reference values ​​for each objective function. To solve The Euclidean distance to its nearest neighbor solution in the high-dimensional target space. For all The average value; HV is the hypervolume index. This is one of the most important indicators for measuring the quality of the solution set in multi-objective optimization. This is a function for calculating the hypervolume. SP is a spacing index used to measure the uniformity of the distribution of solutions within an optimal solution set. A smaller SP value indicates a more uniform distribution of solutions. Efficient solution set The number of solutions.

[0087] The hypervolume index evaluates the overall superiority (convergence) and breadth (diversity) of the solution set. The HV index measures the volume of the space enclosed by the optimal solution set and a reference point. A larger HV value means that the solution set is closer to the ideal optimal surface (good convergence) and covers a wider range of target space (good diversity). The spacing index specifically evaluates the uniformity of the solution set distribution. The SP index calculates the standard deviation of the distance between each solution in the solution set and its nearest neighbor. A smaller SP value means that the solutions are more uniformly distributed on the Pareto front, without clustering. This ensures that the decision-maker faces continuous and uninterrupted choices.

[0088] The quality of the solution set is evaluated based on the hypervolume and spacing indices, and the high-quality solution set with the maximum hypervolume and minimum spacing is selected as the optimal equilibrium solution set. Based on the criteria of "maximizing HV" and "minimizing SP," the final "optimal equilibrium solution set" is selected from the possible efficient solution sets. This ensures that the final set of solutions submitted to the decision-maker possesses the following three characteristics: High quality: Each solution is optimal. Comprehensive: The solution set represents all possible trade-offs. Uniformity: The solutions are evenly distributed, without duplication or missing components, providing a good decision gradient. Automatically eliminating all obviously inferior solutions greatly reduces the decision-maker's screening burden. The subjective "multi-objective trade-off" problem is transformed into an objective "mathematical model screening" process, making the decision results more scientific and reliable. The final output is not a single solution, but a set of optimal trade-off solutions. Decision-makers can make a final selection based on current actual needs and preferences (e.g., choosing the low-cost solution when the budget is tight, or choosing the solution with good coverage when security is emphasized). It is the final step in the entire airport layout optimization process, and its output solution set will be directly used to guide the final airport construction decision.

[0089] Based on the needs and preferences of the actual application scenario, the final airport spatial configuration scheme is selected from the optimal equilibrium solution set. The candidate schemes are ranked through multi-objective optimization, and the optimal solution set (i.e., the optimal equilibrium solution set) is extracted. The airport location, quantity, and service radius configuration of each solution are recorded. Based on the needs and preferences of the actual application scenario (such as cost priority or response speed priority), the final implementation scheme is selected from the solution set, reflecting decision-making flexibility.

[0090] S3. Determine the sequence of inspection tasks based on facility type, history of damage, and urgency of maintenance.

[0091] Specifically, in this embodiment, the inspection tasks are prioritized and the execution order of the drone inspection tasks is determined based on the type of facility, its history of damage, and the urgency of maintenance. This ensures that critical facilities are inspected first, improving maintenance efficiency and responsiveness.

[0092] Step S3, determining the inspection task execution sequence, includes: constructing a task priority assessment model based on facility type weight coefficients, historical severity index of defects, and maintenance urgency rating; using the analytic hierarchy process (AHP) to determine facility type weight coefficients, with structural facilities such as bridges and slopes having higher weights than non-structural facilities such as pavement sections and photovoltaic areas; calculating the historical severity index of defects based on historical inspection data, including quantitative indicators of crack density, deformation rate, and material aging; generating a maintenance urgency rating by combining real-time monitoring data and meteorological forecast information, with facilities that have recently experienced geological disasters or are located in areas affected by severe weather having higher urgency ratings; generating an inspection task priority score by weighted fusion of facility type weights, defect severity, and maintenance urgency ratings; and sorting the inspection tasks according to the priority scores to form an execution sequence, with facilities with higher scores being prioritized for inspection and allocated more drone resources.

[0093] S4. Establish a multi-UAV collaborative trajectory planning model to assign inspection tasks to each UAV and plan conflict-free cyclic inspection paths that meet endurance requirements. Multi-UAV collaborative path planning involves assigning inspection tasks to each UAV and planning conflict-free cyclic inspection paths that meet endurance requirements. This includes constructing a path decision matrix; calculating flight distance and time; and ensuring endurance and safety interval constraints. This enables multi-UAV collaborative, efficient, and safe inspection operations.

[0094] Construct an optimization model under multiple constraints. Its inputs are the set of facilities to be inspected and the fleet of drones. The outputs are the conflict-free mission sequence and detailed flight path of each drone, and ensure that the path meets the physical limitations (endurance) and safety rules (collision avoidance) of the drones.

[0095] Specifically, in this embodiment, a UAV inspection path decision matrix is ​​constructed. , of which elements Indicates drone Are you responsible for the facilities? The inspection task, A value of 0 indicates no responsibility; It is a two-dimensional matrix that defines the task allocation relationship between drones and facilities. The decision matrix definition explicitly specifies the task allocation relationship (i.e., whether drone i is responsible for inspecting facility j). It transforms the abstract "task allocation" into optimizable mathematical variables, providing a decision-making basis for the model. This is a prerequisite for path planning.

[0096] For each drone Generate inspection path sequence ,in For the first in the path One waypoint, This is the length of the path; It is an ordered list that specifies the types of drones. The order of waypoints visited sequentially. The task allocation results are specified as flyable routes, clarifying the number of sorties, inspection order, and flight path for each UAV.

[0097] Calculation path Total flight distance and effective flight time :

[0098] ,

[0099] ,

[0100] in, To account for terrain undulations, the effective distance is calculated. This involves determining the actual flight distance between two waypoints, taking into account terrain undulations (non-linear distances), such as those requiring circling around mountains or climbing. For the segment Meteorological influencing factors; quantified the degree of influence of meteorological conditions (such as headwinds) on flight speed on this flight segment. For the first in the path sequence A waypoint represents the location of a facility that needs to be inspected or a necessary navigation point. For path The length of the path, i.e., the total number of waypoints contained in the path. For path The total flight distance is the sum of the effective distances between all adjacent waypoints along the path. It is a path The effective flight time. This is the estimated time required to complete the flight along this path, taking into account the impact of weather on speed. This is a meteorological sensitivity coefficient used to adjust the intensity of meteorological influence factors. Effective distance and time modeling provides a precise basis for performance evaluation, offering reliable input data for subsequent range constraints and optimization objectives.

[0101] Ensure that the battery life constraints are met:

[0102] ,

[0103] in, For waypoints The time required to conduct inspections. For example, the time required to take photos and collect data. For safety factor ( ), a constant between 0 and 1, is used to reserve a safety margin for the drone's flight time (such as reserving power for return), to prevent the drone from being unable to return due to unexpected circumstances. Drone endurance refers to the maximum continuous flight time a drone can fly on a single charge or refueling. The endurance constraint inequality states that total flight time + total operation time <= safety factor * total endurance. This constraint ensures that the drone has sufficient battery power to safely return to the airport after completing its mission. The endurance constraint guarantee function, a hard constraint condition, establishes a constraint inequality; this is a safety guarantee function that ensures every planned path is executable. It mandates that the drone's "flight time + operation time" must be less than "safety factor × total endurance," reserving a safety margin for return flight and preventing the drone from crashing due to depleted battery power.

[0104] A parallel path planning algorithm based on conflict avoidance is adopted, by introducing a spatiotemporal conflict detection function:

[0105] ,

[0106] in, This represents the spatiotemporal conflict detection function. It is a binary function used to determine whether two drones have clashed at a specific moment. Indicates drone At any moment Position (3D coordinates) (The minimum safe distance, i.e., the straight-line distance). Used to calculate the distance between the two drones at time... The distance between positions.

[0107] The conflict condition indicates that if two drones and At the same time The distance is less than the safe distance Then the function =1 indicates a conflict has occurred; otherwise, it is 0. The conflict detection and avoidance function is a core safety feature that ensures the physical safety of multiple drones operating collaboratively in shared airspace, avoiding the risk of aircraft collisions or signal interference.

[0108] The optimization objective is to minimize the total inspection time.

[0109] ,

[0110] Simultaneously satisfy all conflict constraints The requirement is for all drones... And all the moments The sum of the conflict detection functions must be 0. That is, any conflict must be completely prohibited. The optimization objective is to minimize the maximum task completion time, specifically the time taken for the last drone to complete its inspection task. This objective aims to improve overall inspection efficiency, avoid the "bottleneck effect," and ensure a more balanced workload across all drones. The model ensures that the planned path is not only flyable (meeting endurance requirements) and safe (conflict-free), but also efficient (minimum time). The optimization objective setting function focuses on improving overall operational efficiency. It doesn't aim to reduce the flight time of a single drone, but rather to reduce the time taken for the last drone to complete its task. This avoids the "bottleneck effect," promotes a more balanced task allocation, and shortens the total time of the entire inspection task.

[0111] The above steps provide a complete, reliable, and automated solution for multi-UAV task allocation and path planning. It ensures that the final generated path is not only a "theoretical path," but also an "executable flight plan" that takes into account safety (collision avoidance, endurance), efficiency (shortest time), and practicality (consideration of terrain and weather). This is the key technical guarantee for the entire method to move from theory to practical application.

[0112] S5. Based on real-time environmental changes, a predictive control optimization strategy is employed to continuously update and adjust the trajectory planning model online. Dynamic path replanning is implemented, using predictive control strategies to adjust the UAV path online based on real-time environmental changes (such as weather and airspace variations). This enhances the system's adaptability to dynamic environments and ensures the continuity and safety of inspection missions.

[0113] Specifically, in this embodiment, an environmental state prediction model is constructed within the prediction time domain, based on real-time collected meteorological data. (Meteorological information affecting flight, such as wind speed, precipitation, and visibility, collected at time t), airspace restriction data. (Dynamic changes in airspace such as no-fly zones and restricted flight zones collected at time t) and UAV status data (Based on real-time data collected at time t, including the location, speed, battery level, and health status of each drone), the system predicts environmental change trends over a future period ΔT (the time range for predicting future environmental changes within the MPC framework). This predictive model enables the system to possess "foresight," moving beyond passive response to proactively anticipating risks (such as approaching thunderstorm zones or newly established no-fly zones), providing a basis for proactively adjusting flight paths.

[0114] At every decision moment (The time point at which MPC performs online optimization), using the current UAV state and environmental state as initial conditions, in the prediction time domain Establish a reprogramming optimization model within the future time period covered by the optimization problem:

[0115] ,

[0116] in, These are path weighting coefficients used to adjust the flight path to a specific target point. The importance of the path in the cost function. To comprehensively consider meteorological risks (e.g., flying into thunderstorm areas), airspace restriction risks (e.g., near temporary no-fly zones) and drone status risks A composite risk assessment function (e.g., low battery) for drones. A composite risk assessment function. A function that comprehensively quantifies the risks that drones may face in future flights. These are the weighting coefficients for each risk item, used to adjust the weighting of different risk types (weather, airspace, air condition) in the total risk. The proportion of. For drones The predicted location at time t. This is the decision variable in the optimization model. The location of target point j (usually a facility that needs to be inspected). This is the risk weighting coefficient. It is used to balance the proportion of path cost and overall risk in the objective function.

[0117] Rolling optimization and replanning functionality, model predictive control framework: at each decision time... Starting from the current state, optimize only a finite future time period. The system follows a specific path and executes only the control command for the first step. At the next time step, this process is repeated based on the new measurements. This is an advanced optimization control strategy. Its core advantages are: Feedback correction: Each optimization is based on the latest real state, continuously correcting prediction errors and forming closed-loop control. Rolling implementation: It avoids the impracticality of planning everything at once, making it very suitable for handling dynamically changing environments. Computational feasibility: It decomposes a complex long-term optimization problem into a series of simple short-term optimization problems, ensuring real-time online computation.

[0118] The optimization objective function is: min(path cost + λ * risk cost). The goal is to complete the path as efficiently as possible while minimizing flight risks. This achieves an online trade-off between efficiency and safety. The system can adjust its response based on real-time risks (such as weather risks). airspace risks Body risk Dynamically adjust flight strategies, proactively avoid high-risk areas, and even sacrifice some efficiency for safety when necessary (choosing a longer but safer route).

[0119] Apply the following dynamic constraints:

[0120] ,

[0121] , ,

[0122] in For drones The estimated flight time within the prediction time domain. This is an integral quantity used to calculate the flight time from the current moment to some future moment. For drones At any moment The effective flight speed. This is a quantity that varies over time, reflecting real-time weather conditions. The dynamic impact on speed. A relative velocity term is introduced as the dynamic safety margin factor. This means that the greater the relative speed between the two drones, the greater the minimum safe distance that needs to be maintained, allowing more space and time for maneuvering and avoidance. For drones Cumulative flight time. Used to calculate remaining available flight time. For drones and The relative velocity between them. Vectors, magnitude, and direction all affect the risk of collision.

[0123] Dynamic Constraint 1 (Endurance): Remaining flight time <= Remaining endurance time. This ensures the drone will not exceed its endurance due to replanning. The elapsed time in this constraint is dynamically changing. This ensures the replanned path always considers remaining battery power, and the safety constraint applies throughout the entire process. Dynamic Constraint 2 (Collision Avoidance): Distance between drones >= Static safety distance + Dynamic safety margin. This is a more stringent dynamic collision avoidance constraint. A relative velocity term has been introduced. This means that the greater the relative velocity between the two drones, the larger the safe distance the system reserves for them. This is a very advanced safety feature, providing space and time margin for emergency avoidance.

[0124] A model predictive control framework is used to solve the above optimization problem, and the optimized path for the first time period in the prediction time domain is distributed to each UAV; at the next sampling time... Based on the new state measurements, the above optimization process is repeated to achieve closed-loop feedback control; according to the intensity of environmental changes... Dynamically adjust task priority:

[0125] .

[0126] in, This represents the intensity of environmental change. It is a comprehensive scalar that quantifies the magnitude of change in meteorological and airspace constraint data from the previous moment to the current moment. The initial priority of the task. The static priority determined by step S3 based on facility type, history of damage, etc. This is the adjusted new priority. This is the priority adjustment coefficient. A learning rate parameter greater than 0 controls the sensitivity of priority adjustments to environmental changes. The priority adjustment formula is a heuristic rule. When environmental changes are drastic ( (Large), raising the priority of all tasks, prompting the system to respond to changes more quickly, possibly prioritizing critical tasks or returning to base early to avoid risks.

[0127] S6. Integrate inspection data to generate facility status reports and feed historical performance data back to the airport spatial configuration optimization model and multi-aircraft collaborative trajectory planning model to achieve continuous system self-optimization. Data integration and system self-optimization involve integrating inspection data to generate facility status reports and feeding historical performance data back to the airport layout and route planning models to achieve continuous system optimization. This includes generating health status and urgency reports; calculating historical performance indicators; and dynamically adjusting model parameters (such as weights, service radius, and safety factors). This forms a closed-loop optimization mechanism to improve the overall long-term performance and economy of the system.

[0128] Specifically, in this embodiment, step S6, which integrates inspection data to generate a facility status report and feeds back historical performance data to the airport space configuration optimization model and the multi-aircraft collaborative trajectory planning model, to achieve continuous self-optimization of the system, includes the following steps:

[0129] Collect multi-source inspection data, including facility image data. (High-resolution photos or videos taken by drones for visual inspection of facility surface conditions (such as cracks, damage)) and laser point cloud data. (Three-dimensional spatial data acquired via lidar is used to accurately measure the geometry of the facility (such as deformation and settlement)) and infrared thermal image data. (Temperature distribution data collected by infrared thermal imagers is used to detect internal defects (such as bridge deck voids and photovoltaic panel hot spots)) and drone status data. (Data recorded during the inspection of the drone itself, such as flight path, energy consumption, and flight time, is used to calculate performance indicators.)

[0130] Generate a facility health status assessment report, including a health score (H) and a maintenance urgency index (M).

[0131] ,

[0132] ,

[0133] in For the first The weighting coefficient of each assessment indicator reflects the importance of that indicator in the overall health score. For the first A health score for each assessment indicator, which is a quantitative score of the indicator's status. For the first The urgency of each assessment indicator is quantified by assigning a score to the urgency of the problem it reflects; H is the facility health score, a comprehensive quantitative indicator, with higher values ​​indicating better overall facility condition. M is the maintenance urgency index, a quantitative indicator, with higher values ​​indicating a stronger urgency for facility maintenance. n is the number of assessment indicators. The total number of different dimensions used to assess facility health status (such as crack length, corrosion area, degree of deformation, etc.). The maximum value operation is used to define the maintenance urgency index M. This means that the ultimate urgency of a facility is determined by its most serious problem, consistent with the "weakest link" principle in engineering practice.

[0134] Calculate historical performance metrics, including average inspection interval time. And overall cost-effectiveness ratio E:

[0135] ,

[0136] ,

[0137] in, This represents the average inspection interval. It's an indicator of system response efficiency and inspection frequency. A smaller value indicates more frequent inspections. This represents the total number of facilities. For the first The inspection interval for each target point is the time difference between two consecutive inspections of the facility. E represents the overall cost-effectiveness ratio, a key indicator of system economy, representing the coverage efficiency achieved per unit cost. A higher value indicates better cost-effectiveness of the system. This represents the actual achieved overall coverage effectiveness (calculated from historical data). This represents the actual total lifecycle cost incurred (calculated from historical data).

[0138] Establish a data feedback mechanism to incorporate historical performance data. Input the data into the airport layout multi-objective optimization model and adjust the optimization weight coefficients:

[0139] ,

[0140] ,

[0141] ,

[0142] in Let be the learning rate parameter, and let the weight coefficients satisfy . It is used to control the intensity and direction of the influence of different feedback items on weight adjustment. , , For the first The current weight coefficients of the three objectives of coverage, response, and cost in each optimization cycle. , , This refers to the adjusted weighting coefficients for the next optimization cycle. This represents the change in the maintenance urgency index. The value is positive if the urgency level of the facility increases. This represents the change in the average inspection interval. If the response becomes slower, this value is positive. This represents the change in the health score. If the facility's health deteriorates, the value becomes negative. This represents the change in cost-effectiveness ratio. If economic efficiency deteriorates, this value becomes negative.

[0143] Update the parameters of the multi-aircraft cooperative trajectory planning model, including adjusting the service radius and safety factor:

[0144] ,

[0145] ,

[0146] in, This is the minimum service radius for the current and next cycle. This is the safety factor for the current and next driving cycle. , This is the learning rate parameter. It controls the impact of the feedback term on parameter adjustment.

[0147] Based on the updated model parameters, airport spatial configuration optimization and trajectory planning are re-executed to achieve continuous self-optimization of the system.

[0148] Example 2

[0149] like Figure 2 As shown, this application provides a system architecture diagram for UAV collaborative inspection path planning and airport deployment for highway multi-facility maintenance. It is applied to the UAV collaborative inspection path planning and airport deployment system for highway multi-facility maintenance as described in Embodiment 1, including an environmental characterization system construction module 11, an airport deployment optimization module 12, a task sequence planning module 13, a collaborative path planning module 14, a dynamic replanning module 15, and a closed-loop optimization module 16.

[0150] The environmental characterization system construction module 11 is used to collect data on highway network, facility distribution, topography, airspace restrictions and meteorology, and to construct a multi-dimensional environmental characterization system that includes geospatial constraints, airspace control constraints, meteorological dynamic constraints and facility distribution constraints.

[0151] Airport layout optimization module 12 is used to establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. It uses an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes, and extracts the optimal equilibrium solution set through multi-objective sorting and filtering, thereby determining the final airport spatial configuration scheme.

[0152] The task sequence planning module 13 is used to determine the execution priority sequence of inspection tasks based on facility type, history of damage, and maintenance urgency.

[0153] The collaborative path planning module 14 is used to establish a multi-UAV collaborative trajectory planning model, assign inspection tasks to each UAV, and plan a cyclical inspection path that meets the endurance requirements and is free from conflicts.

[0154] The dynamic replanning module 15 is used to perform online rolling updates and trajectory adjustments on the collaborative trajectory planning model based on real-time environmental changes and using predictive control optimization strategies.

[0155] The closed-loop optimization module 16 is used to integrate inspection data to generate facility status reports and feed back historical performance data to the models of the airport deployment optimization step and the collaborative path planning step, so as to realize the adaptive adjustment and continuous optimization of system parameters.

[0156] Figure 3 This is an electronic device provided in one embodiment of this application. For example... Figure 3 As shown, the electronic device includes at least the following components: processor 101 and memory 100, communication interface 103, and bus 102.

[0157] In this embodiment of the application, memory 100 is used to store executable instructions of processor 101, which, when configured to execute instructions, implements the method as described in the first aspect.

[0158] In embodiments of this application, a computer-readable storage medium includes instructions that instruct a device to perform the method as described in the first aspect. For example, the instructions instruct the device to perform... Figure 1 The method is shown in the process steps.

[0159] In one embodiment of this application, the program operating in the electronic device may be a program that controls a central processing unit (CPU) or similar device to achieve the functions of the above-described embodiments of the present invention (a program that enables the computer to function). Information processed by these systems is then temporarily stored in random access memory (RAM) during processing, and subsequently stored in various ROMs such as read-only memory (FlashROM) and hard disk drives (HDDs), and read, corrected, and written by the CPU as needed.

[0160] It should be noted that a portion of the electronic device described in the above embodiments can also be implemented using a computer. In this case, the program for implementing the control function can be recorded on a computer-readable recording medium, and the program recorded on the recording medium can be read into the computer and executed.

[0161] It should be noted that the computer mentioned here refers to a computer built into an electronic device, employing hardware including an operating system and peripheral devices. Furthermore, computer-readable recording media refers to removable media such as floppy disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage systems such as hard drives built into the computer.

[0162] Furthermore, computer-readable recording media can include: media that dynamically stores programs for short periods of time, such as communication lines used when transmitting programs via networks like the Internet or communication lines like telephone lines; and media that store programs for fixed periods of time, such as volatile memory inside a computer that serves as a server or client in this case. In addition, the aforementioned program can be a program used to implement the above-mentioned functions, or it can be a program that can implement the above-mentioned functions by combining them with programs already recorded in the computer.

[0163] Furthermore, the electronic device in the above embodiments can also be implemented as an assembly (system group) composed of multiple systems. Each system constituting the system group can possess some or all of the functions or functional blocks of the electronic device in the above embodiments. As a system group, it is sufficient to have all the functions or functional blocks of the electronic device.

[0164] Those skilled in the art should recognize that the above embodiments are only used to illustrate this application and are not intended to limit this application. Any appropriate changes and variations made to the above embodiments within the essential spirit and scope of this application fall within the scope of protection claimed in this application.

Claims

1. A method for UAV collaborative inspection path planning and airport deployment for multi-facility maintenance of highways, characterized in that, Includes the following steps: S1: Collect data on highway network, facility distribution, topography, airspace restrictions, and meteorology to construct a multi-dimensional environmental characterization system; S2: Establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. Use an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes. Extract the optimal equilibrium solution set through multi-objective sorting and filtering, and determine the final airport spatial configuration scheme accordingly. S3: Determine the priority sequence for inspection tasks based on facility type, history of damage, and urgency of maintenance; S4: Establish a multi-UAV collaborative trajectory planning model, assign inspection tasks to each UAV, and plan a cyclical inspection path that meets the endurance requirements and is free from conflicts. S5: Based on real-time environmental changes, a predictive control optimization strategy is used to perform online rolling updates and trajectory adjustments on the collaborative trajectory planning model; S6: Integrate inspection data to generate facility status reports, and feed back historical performance data to the models in steps S2 and S4 to achieve adaptive adjustment and continuous optimization of system parameters; S2 includes: Construct a multi-objective optimization function to simultaneously optimize three objectives: coverage, response time, and total cost. An adjustable service radius strategy is adopted, which dynamically adjusts the service radius value within a preset service radius range. The service radius range is calculated and determined based on the UAV's endurance, effective flight speed, and the facility's maximum allowable response time. The airport layout candidate scheme is generated iteratively by combining a heuristic set coverage search algorithm. The candidate solutions are optimized by multi-objective sorting and screening, the optimal equilibrium solution set is extracted, and the corresponding airport location, number and service radius configuration are recorded for each optimal equilibrium solution. Based on the needs and preferences of the actual application scenario, the final airport space configuration scheme is selected from the optimal equilibrium solution set; The planned cyclic inspection path in S4 that meets the range requirements and is conflict-free includes: Construct a drone inspection path decision matrix and define the task allocation relationship between drones and facilities; Generate an inspection path sequence for each drone and determine the order in which waypoints will be visited. Calculate the total flight distance and effective flight time for each path, where the effective flight time takes into account the effects of terrain undulation and weather conditions; To ensure that the endurance constraint is met, the sum of the total flight time and the total inspection operation time shall not exceed the product of the safety factor and the endurance. A parallel path planning algorithm based on conflict avoidance is adopted, and spatiotemporal conflict detection is used to ensure that the minimum safe distance is maintained between UAVs; The optimization objective is to minimize the maximum task completion time, and path planning is performed while satisfying all conflict constraints.

2. The method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways according to claim 1, characterized in that, The construction process of the multidimensional environmental characterization system in S1 includes: The system collects highway network topology, facility spatial distribution, digital elevation model, airspace control information, and multi-dimensional meteorological data through GIS platform, IoT sensors, and meteorological monitoring stations. A multi-dimensional environmental characterization system with constraints is constructed based on the collected multi-source data. By using data fusion technology to perform spatiotemporal registration and consistency processing on multi-source heterogeneous environmental data, and establishing a mapping relationship between environmental data and UAV performance parameters, dynamic updated environmental basic data is provided for airport spatial configuration optimization and flight path planning.

3. The method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways according to claim 1, characterized in that, The construction of the multi-objective optimization function includes: Establish the objective function for maximizing comprehensive coverage effectiveness, expressed as: ， in, Indicates overall coverage effectiveness. Indicates the total number of facilities. Facilities Importance weight, It is the coverage sensitivity coefficient. Indicates drone facilities The coverage quality, It is a drone With facilities The effective distance between them It is the distance attenuation coefficient. For indicator functions, Indicates the minimum coverage quality threshold; The objective function for spatiotemporal response performance is established as follows: ， in, Indicates spatiotemporal response performance. The effective flight speed of the drone, For path Meteorological influencing factors, Meteorological sensitivity coefficient This indicates that all drones The minimum value operation; Establish the life-cycle cost objective function, expressed as: ， in, Total lifecycle cost For the airport Construction costs, Its annual maintenance cost, For drones Purchase cost, For drones Annual energy consumption For the first Annual energy prices, The discount rate is... Project cycle; A decomposition-based multi-objective evolutionary algorithm is employed for collaborative optimization through a comprehensive utility function. ， in, The overall utility value, These are the weighting coefficients. These are the reference maximum values ​​for each objective function.

4. The method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways according to claim 1, characterized in that, The predictive control optimization strategy employed in S5 is used to perform online rolling updates and trajectory adjustments on the cooperative trajectory planning model, including: Construct an environmental state prediction model within the prediction time domain, and predict the environmental change trend in the future time period based on real-time collected meteorological data, airspace restriction data and UAV status data; At each decision point, a replanning optimization model is established within the prediction time domain, using the current state as the initial condition. This model comprehensively considers path cost and compound risk assessment. Apply dynamic constraints, including flight time constraints based on remaining endurance and dynamic collision avoidance constraints that take into account relative speed; A model predictive control framework is used to solve the optimization problem and the optimized path for the first time period is sent to the UAV. The optimization process is repeated based on the new state measurement value at the next sampling time to achieve closed-loop feedback control. The task priority is dynamically adjusted according to the intensity of environmental changes. When environmental changes are drastic, the task priority is increased to enhance the system's responsiveness.

5. The method for collaborative inspection path planning and airport deployment of unmanned aerial vehicles (UAVs) for multi-facility maintenance of highways according to claim 1, characterized in that, The adaptive adjustment and continuous optimization in S6 include: Collect multi-source inspection data, including facility image data, laser point cloud data, infrared thermal map data, and drone status data; Generate a facility health status assessment report, including a weighted summation health score and a repair urgency index based on the most serious problem; Calculate historical performance metrics, including average inspection interval time and overall cost-effectiveness ratio; Establish a data feedback mechanism to input historical performance data into the airport layout multi-objective optimization model, and dynamically adjust the weight coefficients of each objective based on changes in coverage performance, urgency, response time, and health status. Update the parameters of the multi-aircraft cooperative trajectory planning model, including adjusting the minimum service radius based on health status and urgency, and adjusting the endurance safety factor based on response time and cost-effectiveness. Airport spatial configuration optimization and trajectory planning are re-executed based on the updated model parameters to achieve continuous self-optimization of the system.

6. A UAV collaborative inspection path planning and airport deployment system for highway multi-facility maintenance, applied to the UAV collaborative inspection path planning and airport deployment method for highway multi-facility maintenance as described in any one of claims 1 to 5, characterized in that, The system includes: The environmental characterization system construction module is used to collect data on highway network, facility distribution, topography, airspace restrictions, and meteorology to construct a multi-dimensional environmental characterization system. The airport layout optimization module is used to establish a comprehensive optimization model with the objectives of maximizing coverage, minimizing response time, and minimizing total cost. It uses an adjustable service radius strategy and a heuristic set coverage search algorithm to generate candidate airport layout schemes, and extracts the optimal equilibrium solution set through multi-objective sorting and filtering to determine the final airport spatial configuration scheme. The task sequence planning module is used to determine the execution priority sequence of inspection tasks based on facility type, history of damage, and maintenance urgency. The collaborative path planning module is used to establish a multi-UAV collaborative trajectory planning model, assign inspection tasks to each UAV, and plan a cyclical inspection path that meets the endurance requirements and is free from conflicts. The dynamic replanning module is used to perform online rolling updates and trajectory adjustments to the collaborative trajectory planning model based on real-time environmental changes and using predictive control optimization strategies. The closed-loop optimization module is used to integrate inspection data to generate facility status reports and feed back historical performance data to the models in steps S2 and S4 to achieve adaptive adjustment and continuous optimization of system parameters.

Images