A vehicle path planning method based on multi-objective optimization for vehicle-machine cooperative inspection

Abstract

The application belongs to the technical field of intelligent inspection, and provides a vehicle-machine cooperative inspection vehicle path planning method based on multi-objective optimization, which comprises the following steps: establishing a vehicle-machine cooperative inspection path planning model, selecting a minimum set of parking points and planning a vehicle path, completing vehicle scheduling and unmanned aerial vehicle task matching, then for each parking point, generating an initial unmanned aerial vehicle inspection path based on map data using the A-star algorithm, combining visual SLAM to realize environment mapping and real-time positioning, using an improved A-star algorithm and a rolling window mechanism to dynamically avoid obstacles, and in the process of unmanned aerial vehicle flight, fusing multi-source sensor data to estimate the unmanned aerial vehicle attitude position based on a G2O graph optimization framework, combining an adaptive error compensation algorithm to realize robust control of the flight trajectory, and finally judging whether the total length of the inspection path exceeds the single endurance capability of the unmanned aerial vehicle, if yes, segmenting and re-planning the path, and if not, outputting the final path. The application has high inspection efficiency, low cost, stable operation, and is suitable for large-scale tasks.

Description

Technical Field

[0001] This application belongs to the field of intelligent inspection technology, and in particular relates to a vehicle path planning method, device, terminal equipment and storage medium for vehicle-machine collaborative inspection based on multi-objective optimization. Background Technology

[0002] With the rapid growth of renewable energy facilities such as distributed photovoltaic power generation, the workload of their operation and maintenance has increased dramatically. Traditional manual inspection methods are no longer sufficient to meet the requirements of efficient and comprehensive operations. Currently, the power inspection field is gradually adopting intelligent equipment such as drones and inspection robots for automated data collection and using image recognition technology to detect faults. Common inspection methods mainly rely on a single drone or robot to perform fixed-point shooting and data transmission tasks, or adopt automated inspection schemes with pre-set routes.

[0003] However, current methods suffer from limitations in operational scope, insufficient multi-device collaborative scheduling, poor environmental adaptability, and low overall efficiency. Summary of the Invention

[0004] This application provides a vehicle path planning method for vehicle-machine collaborative inspection based on multi-objective optimization, which can solve the problems of limited operating range, insufficient multi-device collaborative scheduling, poor environmental adaptability and low overall efficiency of the current method.

[0005] In a first aspect, embodiments of this application provide a vehicle-to-machine (V2M) collaborative inspection vehicle path planning method based on multi-objective optimization, including: Step 1, establishing a V2M collaborative inspection path planning model, selecting the minimum set of parking points with the premise of covering all poles to be inspected, and minimizing the total travel time of the inspection vehicle as the optimization objective, and generating the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle under multi-objective constraints; Step 2, for each selected parking point, performing initial path planning based on map data containing geospatial information and using the A* algorithm to generate the initial inspection path of the UAV, and combining visual SLAM to achieve environmental mapping and... Real-time positioning, using the improved A-Star algorithm and rolling window mechanism for dynamic obstacle avoidance; Step 3: During the flight of the UAV along the initial inspection path, based on the G2O graph optimization framework, multi-source sensor estimation of UAV attitude and position is integrated, and combined with an adaptive error compensation algorithm, to achieve robust control of flight trajectory in disturbed environments, ensuring operational stability and accuracy; Step 4: Determine whether the total length of the initial inspection path is greater than the single-flight endurance of the UAV. If so, the UAV inspection task is segmented according to a preset distance, and the process returns to Step 2 to replan the path; if not, the initial inspection path is output as the final inspection path.

[0006] In one possible implementation of the first aspect, step 1 above establishes a vehicle-machine collaborative inspection path planning model. This model selects the minimum set of parking points while ensuring coverage of all poles to be inspected. The optimization objective is to minimize the total travel time of the inspection vehicle. Under multiple objective constraints, the model generates the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle. Specifically, this includes:

[0007] Obtain the set of maintenance stations, alternative parking points, and towers to be inspected within the inspection area;

[0008] Let the optimization objective be expressed as:

[0009] ;

[0010] Among them, Z total Z represents the total operating time of the inspection vehicle. q Let q be the working time of the inspection vehicle on the qth working day, where q∈W and W is the set of working days;

[0011] Let the multi-objective constraints include the working time constraint, expressed as:

[0012] ;

[0013] ;

[0014] Where P is the set of parking points, h i The value is 1 if the i-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise. i Let h be the parking time of the inspection vehicle at parking point i, i.e., the inspection time at parking point i, where i∈P, G is the set of vehicle waypoints, G=P∪A, and A is the set of maintenance stations. i’ Indicates the i-th ′ The value is 1 when an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. j’ Indicates the j-th ′ The value is 1 if an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. qi’j’ For the inspection vehicle on the qth working day, from vehicle route point i ′ Driving to vehicle waypoint j ′ The value is 1 if it is true, and 0 otherwise. i’j’ For vehicle path point i ′ and vehicle waypoint j ′ The geographical straight-line distance between them, i ′ ∈G,j ′ ∈G, V is the average speed of the inspection vehicle, Z max The maximum working hours per day for inspection personnel of power companies;

[0015] Let the multi-objective constraint include the inspection vehicle leaving the maintenance station every working day and eventually returning to the maintenance station, represented as:

[0016] ;

[0017] in, The value is 1 if the inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle travels from vehicle waypoint j to vehicle waypoint i on the qth working day, and 0 otherwise.

[0018] Let the multi-objective constraints include that each selected parking spot can only be visited once by the inspection vehicle on a given workday, and that the vehicle must leave the selected parking spot after visiting it. This can be represented as:

[0019] ; ;

[0020] in, The value is 1 if the j-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. For the inspection vehicle on the qth working day, from vehicle route point i ’ The value is 1 when the vehicle reaches path point j, and 0 otherwise. The inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day. ’ The value is 1 if it is true, and 0 otherwise.

[0021] Let the multi-objective constraint include that all poles to be inspected are served by one and only one selected parking point, expressed as:

[0022] ;

[0023] Where, r ti The value is 1 if the t-th pole to be inspected is served by the i-th alternative parking point, and 0 otherwise. T is the set of poles to be inspected.

[0024] Let the multi-objective constraint include that any two endpoints of a vehicle route on a given workday must be visited by the inspection vehicle on that workday, expressed as:

[0025] ;

[0026] in, The value is 1 if the inspection vehicle visits vehicle path point i' on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle visits vehicle path point j' on the qth working day, and 0 otherwise.

[0027] Let the multi-objective constraint include that if a parking spot is visited by an inspection vehicle on a certain workday, then that parking spot must be selected as a parking spot, expressed as:

[0028] ;

[0029] Where, x qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise.

[0030] Assume that the multi-objective constraints include decision variable constraints:

[0031] ;

[0032] ;

[0033] Calculate the parking time C of the inspection vehicle at parking point i. i :

[0034] ;

[0035] Calculate the mission time z of drone k at parking point i. ik :

[0036] ;

[0037] ;

[0038] ;

[0039] Among them, s tk The value d represents 1 if the t-th pole to be inspected is inspected by drone k, and 0 otherwise. ti Let c be the straight-line geographical distance between pole t and parking point i. t U represents the inspection time of the drone at point t on the pole, and U is the drone aggregation time.

[0040] The multi-objective constraints also include that each tower is visited exactly once by a drone, and that at each selected parking spot, the total number of drone takeoffs equals the total number of towers served by that parking spot, expressed as:

[0041] ; ;

[0042] Among them, f ik Let $k$ be the number of times drone $k takes off from parking point $i$.

[0043] Solve the vehicle-machine collaborative inspection path planning model to generate vehicle paths and total inspection vehicle operation time.

[0044] Optionally, in another possible implementation of the first aspect, step 2 above, for each selected parking point, generates an initial inspection path for the UAV based on map data containing geospatial information and using the A* algorithm for initial path planning, specifically including:

[0045] The information on distributed photovoltaic power station inspection objects is structured, and the set of distributed photovoltaic power station objects to be inspected is defined as P. all The distributed photovoltaic power stations to be inspected are respectively denoted as P1, P2, ..., P... n Let P be the elevation of the highest point where the nth distributed photovoltaic power station module to be inspected is installed. n h;

[0046] Local target location and marking will be performed on the distributed photovoltaic power station to be inspected, and the data will be collected in set P. all One or more distributed photovoltaic power stations to be inspected that are within a straight-line distance of less than a set value X are classified as a local target, denoted as A. n And measure the latitude and longitude information of the outer edge feature points of the local target;

[0047] According to set P all The straight-line distance between any two distributed photovoltaic power station objects to be inspected is used to determine the range of values ​​for the set value X. The minimum value of the set value X is the minimum straight-line distance between any two objects in the set, and the maximum value is the maximum straight-line distance between any two objects.

[0048] Determine whether the outer edge feature point of the local target meets the following conditions: it belongs to the outer edge of the photovoltaic module area, the outer edge angle is not 180°, and there is at least one straight line passing through the outer edge feature point that does not intersect with the photovoltaic module in the local target.

[0049] Among the outer edge feature points that meet the conditions, the line connecting the two outer edge feature points that are farthest apart is selected as a diagonal of the local target rectangular inspection area, denoted as L1. Another diagonal L2, of the same length as diagonal L1 and coinciding with its center point, is constructed, and the angle between the two diagonals is adjusted to form a rectangular inspection area containing all outer edge feature points within the local target, denoted as S. n ;

[0050] Based on the size and focal length of the charge-coupled device (CCD) of the camera mounted on the drone and the working altitude L of the drone, the inspection angle of the drone is determined, and the imaging area of ​​a single shot is determined by the angle.

[0051] The starting point for the UAV inspection of the nth local target is determined as S. n s, The inspection endpoint is S n e and the highest elevation of the component installation point within the local target S n h, Sn h represents the maximum value of the installation elevation of each distributed photovoltaic power station component to be inspected within the local target area;

[0052] Local target rectangular inspection area S n The vertex is used as the inspection starting point, and the long side of the rectangle connected to the inspection starting point is used as the initial inspection route. The S-shaped inspection method is used to plan the path of the rectangular inspection area. The latitude and longitude information of the UAV inspection path is calculated and the initial inspection path of the UAV in the local target is generated by using the imaging area stitching to cover the local target as a constraint.

[0053] After completing the initial inspection path planning for each local target, the straight-line distance between different local targets is calculated based on the latitude and longitude information of each local target. The Traveling Salesman Problem (TSP) is used to determine the shortest inspection path for the UAV to traverse each local target and return to the starting local target, thus forming the initial inspection path between multiple local targets.

[0054] Optionally, in another possible implementation of the first aspect, step 2 above, which combines visual SLAM to achieve environmental mapping and real-time localization, specifically includes:

[0055] During the flight of the UAV along the initial inspection path, continuous environmental image data is collected using airborne visual sensors.

[0056] Feature extraction and feature matching are performed on environmental image data, and the relative motion information of the UAV during flight is estimated based on the feature correlation between adjacent image frames;

[0057] Based on relative motion information, the current pose of the UAV is updated in real time, and the UAV pose is fused with environmental feature information to construct a local map of the UAV's flight environment.

[0058] During the continuous flight of the drone, the local map is updated and optimized in real time, and the drone's real-time pose information is output.

[0059] Optionally, in another possible implementation of the first aspect, step 2 above, which utilizes the improved A* algorithm and the scrolling window mechanism for dynamic obstacle avoidance, specifically includes:

[0060] During the flight of the UAV along the initial inspection path, the local environment map constructed based on visual SLAM and the real-time pose information of the UAV are used to obtain information on environmental changes ahead of the UAV's flight path in real time.

[0061] In response to environmental change information, an environmental risk cost reflecting obstacle distribution and safety distance constraints is introduced on the path cost function of the A-Star algorithm to dynamically evaluate the local cost of the UAV flight path and form an improved path search model.

[0062] Starting from the current pose of the drone, the drone's flight path is locally replanned within the local space covered by the preset scrolling window to generate a local flight path that avoids dynamic obstacles.

[0063] As the drone's pose information is updated in real time during flight, the position of the scrolling window is dynamically updated, and the local path replanning process is repeatedly executed. This enables continuous dynamic obstacle avoidance during the drone inspection process without changing the overall inspection sequence, so as to output a safe and feasible inspection flight path.

[0064] Optionally, in another possible implementation of the first aspect, step 3 above, during the flight of the UAV along the initial inspection path, uses a G2O graph optimization framework to fuse multi-source sensor estimation of the UAV's attitude and position, and combines it with an adaptive error compensation algorithm to achieve robust flight trajectory control under disturbed environments, ensuring operational stability and accuracy. Specifically, this includes:

[0065] The attitude and position estimation of a UAV is based on the G2O graph optimization framework, which integrates multi-source sensor data. This includes: using the FAST feature point detection algorithm to extract feature points from the image data acquired by the UAV to achieve visual positioning of the UAV; combining acceleration and angular velocity data acquired by MEMS inertial devices, latitude and longitude positioning information provided by GPS, and altitude positioning information acquired by barometers, and introducing extended Kalman filtering under the G2O graph-based nonlinear optimization framework to jointly optimize the multi-source sensor data and obtain the attitude and position estimation results of the UAV.

[0066] Based on the attitude and position estimation results, an adaptive law is adopted for robust attitude angle tracking control of the UAV flight trajectory. Specifically, for the time delay function... and According to the Lyapunove functional, the stable solution of the flight trajectory robust tracking control satisfies:

[0067] ;

[0068] Among them, h1 and h2 are normal values;

[0069] By introducing a yaw and pitch oscillation vector, the reference motion equation for UAV flight trajectory tracking is controlled in a closed loop, and the asymptotic stability of UAV flight trajectory tracking control is achieved based on the Lyapunove stability principle.

[0070] Based on the adaptive fusion of multi-sensor information, the flight trajectory tracking error is compensated in real time, realizing flight trajectory tracking control and error correction compensation of UAV under uncertain airflow disturbance.

[0071] Secondly, embodiments of this application provide a vehicle-to-machine (V2M) collaborative inspection vehicle path planning device based on multi-objective optimization, comprising: a generation module for establishing a V2M collaborative inspection path planning model, selecting the minimum set of parking points with the premise of covering all poles to be inspected, and minimizing the total travel time of the inspection vehicle as the optimization objective, and generating a vehicle path, a set of selected parking points, and a total operation time of the inspection vehicle under multi-objective constraints; and a path planning module for generating an initial inspection path for the UAV based on map data containing geospatial information and using the A* algorithm for each selected parking point, and combining visual SLAM to realize environmental mapping. With real-time positioning, dynamic obstacle avoidance is achieved using an improved A-satellite algorithm and a rolling window mechanism. The control module, based on the G2O graph optimization framework, integrates multi-source sensor estimation to determine the UAV's attitude and position during its flight along the initial inspection path. Combined with an adaptive error compensation algorithm, this enables robust flight trajectory control under disturbed environments, ensuring operational stability and accuracy. The judgment and output module determines whether the total length of the initial inspection path exceeds the UAV's single-flight endurance. If so, the UAV inspection task is segmented according to a preset distance, and the path is replanned in the path planning module. If not, the initial inspection path is output as the final inspection path.

[0072] Thirdly, embodiments of this application provide a terminal device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a vehicle path planning method for vehicle-machine collaborative inspection based on multi-objective optimization as described above.

[0073] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the aforementioned vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization.

[0074] Beneficial Effects: By establishing a multi-objective optimized vehicle-to-machine (V2M) collaborative inspection path planning model, integrated collaborative scheduling of inspection vehicle parking point selection, vehicle routes, and UAV tasks was achieved, significantly reducing vehicle travel distance and ineffective waiting time, and improving overall operational efficiency. Combining high-precision maps, improved A-satellite algorithms, and visual SLAM technology enhanced the UAV's real-time positioning and dynamic obstacle avoidance capabilities in complex environments, ensuring the safety and feasibility of flight paths. Employing a G2O graph optimization framework to fuse multi-source sensor data, combined with adaptive error compensation control, improved the UAV's flight stability and trajectory tracking accuracy in disturbed environments. Furthermore, through endurance assessment and task segmentation mechanisms, the system ensures the complete and reliable execution of large-scale inspection tasks even under UAV range limitations, demonstrating good adaptability, robustness, and scalability. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 This is a flowchart illustrating a vehicle path planning method for vehicle-machine collaborative inspection based on multi-objective optimization, provided in an embodiment of this application.

[0077] Figure 2 This is a schematic diagram of the outer edge feature points and the local target rectangular region provided in an embodiment of this application;

[0078] Figure 3 This is a schematic diagram of the "S"-shaped route planning for a local target rectangular area inspection, provided in one embodiment of this application.

[0079] Figure 4 This is a schematic diagram of the lens imaging principle provided in an embodiment of this application;

[0080] Figure 5 This is a flowchart of an embodiment of the obstacle avoidance function of a drone provided in this application;

[0081] Figure 6 This is a schematic diagram of the structure of a vehicle-machine collaborative inspection vehicle path planning device based on multi-objective optimization, provided in an embodiment of this application;

[0082] Figure 7 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this application. Detailed Implementation

[0083] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0084] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0085] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0086] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0087] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0088] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0089] The following is a detailed description of a vehicle path planning method, device, terminal equipment, and storage medium based on multi-objective optimization for vehicle-machine collaborative inspection, provided in this application, with reference to the accompanying drawings.

[0090] Figure 1 The illustration shows a flowchart of a vehicle path planning method for vehicle-machine collaborative inspection based on multi-objective optimization provided in an embodiment of this application.

[0091] like Figure 1 As shown, the vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization includes the following steps:

[0092] Step 1: Establish a vehicle-machine collaborative inspection path planning model. Select the minimum set of parking points with the premise of covering all poles to be inspected, and take minimizing the total driving time of the inspection vehicle as the optimization objective. Under multiple objective constraints, generate the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle.

[0093] Furthermore, in this embodiment of the application, step 1 above includes:

[0094] Obtain the set of maintenance stations, alternative parking points, and towers to be inspected within the inspection area;

[0095] Let the optimization objective be expressed as:

[0096] ;

[0097] Among them, Z total Z represents the total operating time of the inspection vehicle. q Let q be the working time of the inspection vehicle on the qth working day, where q∈W and W is the set of working days;

[0098] Let the multi-objective constraints include the working time constraint, expressed as:

[0099] ;

[0100] ;

[0101] Where P is the set of parking points, h i The value is 1 if the i-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise. i Let h be the parking time of the inspection vehicle at parking point i, i.e., the inspection time at parking point i, where i∈P, G is the set of vehicle waypoints, G=P∪A, and A is the set of maintenance stations. i’ Indicates the i-th ′ The value is 1 when an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. j’ Indicates the j-th ′ The value is 1 if an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. qi’j’ For the inspection vehicle on the qth working day, from vehicle route point i ′ Driving to vehicle waypoint j ′ The value is 1 if it is true, and 0 otherwise. i’j’ For vehicle path point i ′ and vehicle waypoint j ′ The geographical straight-line distance between them, i ′ ∈G,j ′ ∈G, V is the average speed of the inspection vehicle, Z max The maximum working hours per day for inspection personnel of power companies;

[0102] Let the multi-objective constraint include the inspection vehicle leaving the maintenance station every working day and eventually returning to the maintenance station, represented as:

[0103] ;

[0104] in, The value is 1 if the inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle travels from vehicle waypoint j to vehicle waypoint i on the qth working day, and 0 otherwise.

[0105] Let the multi-objective constraints include that each selected parking spot can only be visited once by the inspection vehicle on a given workday, and that the vehicle must leave the selected parking spot after visiting it. This can be represented as:

[0106] ; ;

[0107] in, The value is 1 if the j-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. For the inspection vehicle on the qth working day, from vehicle route point i ’ The value is 1 when the vehicle reaches path point j, and 0 otherwise. The inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day. ’ The value is 1 if it is true, and 0 otherwise.

[0108] Let the multi-objective constraint include that all poles to be inspected are served by one and only one selected parking point, expressed as:

[0109] ;

[0110] Where, r ti The value is 1 if the t-th pole to be inspected is served by the i-th alternative parking point, and 0 otherwise. T is the set of poles to be inspected.

[0111] Let the multi-objective constraint include that any two endpoints of a vehicle route on a given workday must be visited by the inspection vehicle on that workday, expressed as:

[0112] ;

[0113] in, The value is 1 if the inspection vehicle visits vehicle path point i' on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle visits vehicle path point j' on the qth working day, and 0 otherwise.

[0114] Let the multi-objective constraint include that if a parking spot is visited by an inspection vehicle on a certain workday, then that parking spot must be selected as a parking spot, expressed as:

[0115] ;

[0116] Where, x qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise.

[0117] Assume that the multi-objective constraints include decision variable constraints:

[0118] ;

[0119] ;

[0120] Calculate the parking time C of the inspection vehicle at parking point i. i :

[0121] ;

[0122] It should be noted that, since the inspection vehicle needs to launch its onboard drones at each parking point to complete the inspection task, and all drones operate simultaneously, the inspection vehicle will only transport all drones to the next parking point once all drones have completed their respective inspection tasks and landed back on the vehicle. Therefore, the parking time of the inspection vehicle at each selected parking point is equal to the task time of the drone that takes the longest at that parking point.

[0123] Calculate the mission time z of drone k at parking point i. ik :

[0124] ;

[0125] ;

[0126] ;

[0127] Among them, s tk The value d represents 1 if the t-th pole to be inspected is inspected by drone k, and 0 otherwise. ti Let c be the straight-line geographical distance between pole t and parking point i. t U represents the inspection time of the drone at point t on the pole, and U is the drone aggregation time.

[0128] It should be noted that at each selected parking point, the mission time for each drone consists of two parts: the flight time between the parking point and the pole, and the inspection time at the pole.

[0129] The multi-objective constraints also include that each tower is visited exactly once by a drone, and that at each selected parking spot, the total number of drone takeoffs equals the total number of towers served by that parking spot, expressed as:

[0130] ; ;

[0131] Among them, f ik Let $k$ be the number of times drone $k takes off from parking point $i$.

[0132] Solve the vehicle-machine collaborative inspection path planning model to generate vehicle paths and total inspection vehicle operation time.

[0133] Step 2: For each selected parking point, based on map data containing geospatial information and using the AlphaGo algorithm, perform initial path planning to generate the initial inspection path for the UAV. Combine visual SLAM to achieve environmental mapping and real-time positioning, and use the improved AlphaGo algorithm and scrolling window mechanism for dynamic obstacle avoidance.

[0134] Here, A* algorithm refers to a commonly used pathfinding and graph traversal algorithm. In A*, it starts by checking the adjacent squares of the starting point and then expands outwards until the target is found. The specific steps include:

[0135] (1) Start from point A and add it to an open list of squares. This open list is similar to a shopping list. Currently, the open list contains only one item, point A, and more items will be added gradually. The squares in the open list represent paths that may or may not be traversed. Essentially, the open list is a list of squares to be checked.

[0136] (2) Examine the squares adjacent to the starting point A (ignore squares occupied by walls, rivers, and other illegal terrain), and add the walkable or reachable squares to the open list. Set the starting point A as the parent node or parent square of these squares;

[0137] (3) Remove A from the open list and add it to the closed list. Each cell in the closed list is one that no longer needs to be considered.

[0138] Furthermore, in this embodiment of the application, step 2 above, which involves generating an initial inspection path for the UAV for each selected parking point based on map data containing geospatial information and using the A* algorithm, specifically includes:

[0139] The information on distributed photovoltaic power station inspection objects is structured, and the set of distributed photovoltaic power station objects to be inspected is defined as P. all The distributed photovoltaic power stations to be inspected are respectively denoted as P1, P2, ..., P...n Let P be the elevation of the highest point where the nth distributed photovoltaic power station module to be inspected is installed. n h;

[0140] Local target location and marking will be performed on the distributed photovoltaic power station to be inspected, and the data will be collected in set P. all One or more distributed photovoltaic power stations to be inspected that are within a straight-line distance of less than a set value X are classified as a local target, denoted as A. n And measure the latitude and longitude information of the outer edge feature points of the local target;

[0141] According to set P all The straight-line distance between any two distributed photovoltaic power station objects to be inspected is used to determine the range of values ​​for the set value X. The minimum value of the set value X is the minimum straight-line distance between any two objects in the set, and the maximum value is the maximum straight-line distance between any two objects.

[0142] The determination is made based on whether an outer edge feature point in a local target meets the following conditions: it belongs to the outer edge of the photovoltaic module area, the outer edge angle is not 180°, and there exists at least one straight line passing through the outer edge feature point that does not intersect with the photovoltaic modules within the local target. In one embodiment, a schematic diagram of the outer edge feature point is shown below. Figure 2 As shown.

[0143] Among the outer edge feature points that meet the conditions, the line connecting the two outer edge feature points that are farthest apart is selected as a diagonal of the local target rectangular inspection area, denoted as L1. Another diagonal L2, of the same length as diagonal L1 and coinciding with its center point, is constructed, and the angle between the two diagonals is adjusted to form a rectangular inspection area containing all outer edge feature points within the local target, denoted as S. n ;

[0144] Based on the size and focal length of the charge-coupled device (CCD) of the camera mounted on the drone and the working altitude L of the drone, the inspection angle of the drone is determined, and the imaging area of ​​a single shot is determined by the angle.

[0145] The starting point for the UAV inspection of the nth local target is determined as S. n s, The inspection endpoint is S n e and the highest elevation of the component installation point within the local target S n h, S n h represents the maximum value of the installation elevation of each distributed photovoltaic power station component to be inspected within the local target area;

[0146] like Figure 3 As shown, the rectangular inspection area S of the local target is... nThe vertex is used as the inspection starting point, and the long side of the rectangle connected to the inspection starting point is used as the initial inspection route. The S-shaped inspection method is used to plan the path of the rectangular inspection area. The latitude and longitude information of the UAV inspection path is calculated and the initial inspection path of the UAV in the local target is generated by using the imaging area stitching to cover the local target as a constraint.

[0147] After completing the initial inspection path planning for each local target, the straight-line distance between different local targets is calculated based on the latitude and longitude information of each local target. The Traveling Salesman Problem (TSP) is used to determine the shortest inspection path for the UAV to traverse each local target and return to the starting local target, thus forming the initial inspection path between multiple local targets.

[0148] In one embodiment, such as Figure 4 As shown, the angle of view is determined by the size of the camera's charge-coupled device (CCD), focal length, and working height. The working height is denoted as L. According to the imaging principle of a convex lens, the working height : angle of view = focal length : size of the camera's charge-coupled device (CCD). The area of ​​the imaging region is determined by the angle of view, that is, the imaging region area = angle of view (length) × angle of view (width).

[0149] Furthermore, in this embodiment, step 2 above, which combines visual SLAM to achieve environmental mapping and real-time localization, specifically includes:

[0150] During the flight of the UAV along the initial inspection path, continuous environmental image data is collected using airborne visual sensors.

[0151] Feature extraction and feature matching are performed on environmental image data, and the relative motion information of the UAV during flight is estimated based on the feature correlation between adjacent image frames;

[0152] Based on relative motion information, the current pose of the UAV is updated in real time, and the UAV pose is fused with environmental feature information to construct a local map of the UAV's flight environment.

[0153] During the continuous flight of the drone, the local map is updated and optimized in real time, and the drone's real-time pose information is output.

[0154] Furthermore, in the embodiments of this application, such as Figure 5 As shown, step 2 above, which utilizes the improved A* algorithm and the scrolling window mechanism for dynamic obstacle avoidance, specifically includes:

[0155] During the flight of the UAV along the initial inspection path, the local environment map constructed based on visual SLAM and the real-time pose information of the UAV are used to obtain information on environmental changes ahead of the UAV's flight path in real time.

[0156] In response to environmental change information, an environmental risk cost reflecting obstacle distribution and safety distance constraints is introduced on the path cost function of the A-Star algorithm to dynamically evaluate the local cost of the UAV flight path and form an improved path search model.

[0157] Starting from the current pose of the drone, the drone's flight path is locally replanned within the local space covered by the preset scrolling window to generate a local flight path that avoids dynamic obstacles.

[0158] As the drone's pose information is updated in real time during flight, the position of the scrolling window is dynamically updated, and the local path replanning process is repeatedly executed. This enables continuous dynamic obstacle avoidance during the drone inspection process without changing the overall inspection sequence, so as to output a safe and feasible inspection flight path.

[0159] Step 3: During the flight of the UAV along the initial inspection path, the attitude and position of the UAV are estimated by integrating multi-source sensors based on the G2O graph optimization framework and combined with the adaptive error compensation algorithm to achieve robust control of the flight trajectory in a disturbed environment, ensuring the stability and accuracy of the operation.

[0160] Furthermore, in this embodiment of the application, step 3 above includes:

[0161] The attitude and position estimation of a UAV is based on the G2O graph optimization framework, which integrates multi-source sensor data. This includes: using the FAST feature point detection algorithm to extract feature points from the image data acquired by the UAV to achieve visual positioning of the UAV; combining acceleration and angular velocity data acquired by MEMS inertial devices, latitude and longitude positioning information provided by GPS, and altitude positioning information acquired by barometers, and introducing extended Kalman filtering under the G2O graph-based nonlinear optimization framework to jointly optimize the multi-source sensor data and obtain the attitude and position estimation results of the UAV.

[0162] Based on the attitude and position estimation results, an adaptive law is adopted for robust attitude angle tracking control of the UAV flight trajectory. Specifically, for the time delay function... and According to the Lyapunove functional, the stable solution of the flight trajectory robust tracking control satisfies:

[0163] ;

[0164] Among them, h1 and h2 are normal values;

[0165] By introducing a yaw and pitch oscillation vector, the reference motion equation for UAV flight trajectory tracking is controlled in a closed loop, and the asymptotic stability of UAV flight trajectory tracking control is achieved based on the Lyapunove stability principle.

[0166] Based on the adaptive fusion of multi-sensor information, the flight trajectory tracking error is compensated in real time, realizing flight trajectory tracking control and error correction compensation of UAV under uncertain airflow disturbance.

[0167] Step 4: Determine whether the total length of the initial inspection path is greater than the single-use endurance of the drone. If so, divide the drone inspection task into segments according to the preset distance and return to Step 2 to replan the path; otherwise, output the initial inspection path as the final inspection path.

[0168] This application provides a vehicle-to-machine (V2M) collaborative inspection vehicle path planning method based on multi-objective optimization. By establishing a multi-objective optimized V2M collaborative inspection path planning model, it achieves integrated collaborative scheduling of inspection vehicle parking point selection, vehicle path, and UAV missions, significantly reducing vehicle travel distance and ineffective waiting time, and improving overall operational efficiency. Combining high-precision maps, an improved A-satellite algorithm, and visual SLAM technology enhances the real-time positioning and dynamic obstacle avoidance capabilities of UAVs in complex environments, ensuring the safety and feasibility of flight paths. Employing a G2O graph optimization framework to fuse multi-source sensor data, combined with adaptive error compensation control, improves the flight stability and trajectory tracking accuracy of UAVs in disturbed environments. Furthermore, through endurance judgment and task segmentation mechanisms, it ensures that large-scale inspection tasks can still be executed completely and reliably under UAV range limitations. The system as a whole possesses good adaptability, robustness, and scalability.

[0169] The following specific embodiment illustrates the solution of this application.

[0170] This embodiment uses a real-world distributed power line inspection task as a background. The relevant experimental parameters are as follows: one inspection vehicle, an average travel speed of 40 km / h, a single parking radius of 2.5 km, an average flight speed of the UAV between the power pole and the parking point of 60 km / h, and an inspection time of 20 minutes per power pole. The objective is to minimize the total vehicle travel time and complete a comprehensive inspection of all power poles while meeting the constraints of the operating radius and range.

[0171] In the path planning phase, a coverage optimization model is first constructed based on pole distribution data. Under the constraints of two objective functions—minimum number of parking points and minimum vehicle travel distance—the optimal combination of parking points is solved, and vehicle-machine task matching and scheduling path generation are completed. For the path generation and obstacle avoidance problems of the UAV, the system uses a SLAM algorithm based on LiDAR input for mapping and localization. After the UAV starts its LiDAR device, the radar performs a 360° rotation scan of the surrounding environment, constructing a grid-based 2D map. Due to Gaussian noise in the radar input, a Kalman filter algorithm is introduced to recursively estimate and filter the attitude data, improving positioning accuracy and mapping stability.

[0172] The optimal driving route for the inspection vehicle is calculated using a vehicle-to-machine (V2X) collaborative inspection vehicle route planning method. It takes the inspection vehicle three working days to complete the entire maintenance area. The driving routes for each working day are as follows:

[0173] Weekday 1: Maintenance Station — Alternate Parking Point 4 — Alternate Parking Point 19 — Alternate Parking Point 22 — Alternate Parking Point 20 — Alternate Parking Point 9 — Alternate Parking Point 12 — Alternate Parking Point 5 — Maintenance Station;

[0174] Weekday 2: Maintenance Station — Alternate Parking Point 2 — Alternate Parking Point 1 — Alternate Parking Point 7 — Alternate Parking Point 24 — Alternate Parking Point 25 — Alternate Parking Point 6 — Maintenance Station;

[0175] Workday 3: Maintenance Station — Alternate Parking Point 3 — Alternate Parking Point 17 — Alternate Parking Point 23 — Maintenance Station.

[0176] The operating hours of the inspection vehicle on each workday are as follows: Workday 1: 459.39 min; Workday 2: 448.71 min; Workday 3: 449.35 min.

[0177] During operation, the inspection vehicle uses a G2O graph optimization algorithm to fuse information from multiple sources, including vision, inertial, and GPS, to achieve high-frequency attitude estimation. Combined with an adaptive error compensation strategy, it achieves robust control of the flight trajectory under disturbed environments. The system can compare the real-time path length with the UAV's endurance. If the total path exceeds the range limit, it automatically divides the task into multiple sub-task segments and iteratively replans the path; otherwise, it directly outputs the flight path and executes the task.

[0178] Finally, the robot's navigation function is configured using a navigation package based on the ROS platform. Multiple sensor drive nodes, including laser sensors, IMUs, and odometry, are started by configuring the launch file, and coordinate transformation relationships are established. The system employs two cost map structures for global path planning and local path planning respectively, supporting real-time obstacle avoidance. After inputting the target point, the navigation module drives the autonomous vehicle platform to move along the planned path, while simultaneously providing graphical monitoring on a PC.

[0179] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0180] Corresponding to the above embodiment, a vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization, Figure 6The diagram shows a structural block diagram of a vehicle-machine collaborative inspection vehicle path planning device based on multi-objective optimization according to an embodiment of this application. For ease of explanation, only the parts related to the embodiment of this application are shown.

[0181] Reference Figure 6 The device 600 includes:

[0182] The generation module 601 is used to establish a vehicle-machine collaborative inspection path planning model, select the minimum set of parking points on the premise of covering all poles to be inspected, and take minimizing the total driving time of the inspection vehicle as the optimization objective. Under multiple objective constraints, it generates the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle.

[0183] The path planning module 602 is used to perform initial path planning for each selected parking point based on map data containing geospatial information and using the A-Star algorithm to generate the initial inspection path of the UAV. It combines visual SLAM to realize environmental mapping and real-time positioning, and uses the improved A-Star algorithm and scrolling window mechanism for dynamic obstacle avoidance.

[0184] The control module 603 is used to estimate the attitude and position of the UAV based on the G2O graph optimization framework and integrate multi-source sensors during the flight of the UAV along the initial inspection path, and combine it with an adaptive error compensation algorithm to achieve robust flight trajectory control in a disturbed environment, ensuring operational stability and accuracy.

[0185] The judgment output module 604 is used to determine whether the total length of the initial inspection path is greater than the single-use endurance of the drone. If so, the drone inspection task is divided into segments according to the preset distance and returned to the path planning module to replan the path; if not, the initial inspection path is output as the final inspection path.

[0186] In practical use, the vehicle-machine collaborative inspection vehicle path planning device based on multi-objective optimization provided in this application embodiment can be configured in any terminal device to execute the aforementioned vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization.

[0187] This application provides a vehicle-to-machine (V2M) collaborative inspection vehicle path planning device based on multi-objective optimization. By establishing a multi-objective optimized V2M collaborative inspection path planning model, it achieves integrated collaborative scheduling of inspection vehicle parking point selection, vehicle path, and UAV missions, significantly reducing vehicle travel distance and ineffective waiting time, and improving overall operational efficiency. Combining high-precision maps, an improved A-satellite algorithm, and visual SLAM technology, it enhances the real-time positioning and dynamic obstacle avoidance capabilities of UAVs in complex environments, ensuring the safety and feasibility of flight paths. Employing a G2O graph optimization framework to fuse multi-source sensor data, combined with adaptive error compensation control, it improves the flight stability and trajectory tracking accuracy of UAVs in disturbed environments. Furthermore, through endurance judgment and task segmentation mechanisms, it ensures that large-scale inspection tasks can still be executed completely and reliably even under UAV range limitations. The system as a whole possesses good adaptability, robustness, and scalability.

[0188] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0189] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0190] To implement the above embodiments, this application also proposes a terminal device.

[0191] Figure 7 This is a schematic diagram of the structure of a terminal device according to an embodiment of this application.

[0192] like Figure 7 As shown, the terminal device 200 includes:

[0193] The system includes a memory 210 and at least one processor 220, and a bus 230 connecting different components (including the memory 210 and the processor 220). The memory 210 stores a computer program, which, when executed by the processor 220, implements a vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization as described in the embodiments of this application.

[0194] Bus 230 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0195] Terminal device 200 typically includes various electronically readable media. These media can be any available media that can be accessed by terminal device 200, including volatile and non-volatile media, removable and non-removable media.

[0196] Memory 210 may also include computer system readable media in the form of volatile memory, such as RAM 240 and / or cache 250. Terminal device 200 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 260 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 7 Not shown; usually referred to as a "hard drive"). Although Figure 7 As not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 230 via one or more data media interfaces. Memory 210 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this application.

[0197] A program / utility 280 having a set (at least one) of program modules 270 may be stored, for example, in memory 210. Such program modules 270 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 270 typically perform the functions and / or methods described in the embodiments of this application.

[0198] Terminal device 200 can also communicate with one or more external devices 290 (e.g., keyboard, pointing device, display 291, etc.), and with one or more devices that enable a user to interact with terminal device 200, and / or with any device that enables terminal device 200 to communicate with one or more other computing devices (e.g., network card, modem). This communication can be performed via input / output (I / O) interface 292. Furthermore, terminal device 200 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 293. As shown, network adapter 293 communicates with other modules of terminal device 200 via bus 230. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with terminal device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0199] The processor 220 performs various functional applications and data processing by running programs stored in the memory 210.

[0200] It should be noted that the implementation process and technical principles of the terminal device in this embodiment are explained in the foregoing description of a vehicle-machine collaborative inspection vehicle path planning method based on multi-objective optimization in this application embodiment, and will not be repeated here.

[0201] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0202] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the above-described method embodiments.

[0203] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some cases, the computer-readable medium cannot be an electrical carrier signal or a telecommunication signal.

[0204] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0205] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0206] In the embodiments provided in this application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0207] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0208] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A vehicle path planning method for vehicle-machine collaborative inspection based on multi-objective optimization, characterized in that, include: Step 1: Establish a vehicle-machine collaborative inspection path planning model. Select the minimum set of parking points with the premise of covering all poles to be inspected, and take minimizing the total driving time of the inspection vehicle as the optimization objective. Under multiple objective constraints, generate the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle. Step 2: For each selected parking point, based on map data containing geospatial information and using the AlphaGo algorithm, perform initial path planning to generate the initial inspection path for the UAV. Combine visual SLAM to achieve environmental mapping and real-time positioning, and use the improved AlphaGo algorithm and scrolling window mechanism for dynamic obstacle avoidance. The dynamic obstacle avoidance using the improved A* algorithm and the scrolling window mechanism specifically includes: During the flight of the UAV along the initial inspection path, the local environment map constructed based on visual SLAM and the real-time pose information of the UAV are used to obtain information on environmental changes ahead of the UAV's flight path in real time. Based on the environmental change information, an environmental risk cost reflecting obstacle distribution and safe distance constraints is introduced into the path cost function of the A-Star algorithm to dynamically evaluate the local cost of the UAV flight path and form an improved path search model. Starting from the current pose of the drone, the drone's flight path is locally replanned within the local space covered by the preset scrolling window to generate a local flight path that avoids dynamic obstacles. As the pose information of the UAV is updated in real time during flight, the position of the scrolling window is dynamically updated, and the local path replanning process is repeatedly executed. In this way, without changing the overall inspection sequence, continuous dynamic obstacle avoidance of the UAV during the inspection process is achieved, so as to output a safe and feasible inspection flight path. Step 3: During the UAV's flight along the initial inspection path, based on the G2O graph optimization framework, multi-source sensor data is fused to estimate the UAV's attitude and position. Combined with an adaptive error compensation algorithm, robust flight trajectory control under disturbed environments is achieved, ensuring operational stability and accuracy. The estimation of the UAV's attitude and position based on the G2O graph optimization framework and multi-source sensor data includes: The FAST feature point detection algorithm is used to extract feature points from the image data acquired by the UAV to achieve visual positioning of the UAV. The acceleration and angular velocity data acquired by MEMS inertial devices, the latitude and longitude positioning information provided by GPS, and the altitude positioning information acquired by barometer are combined. Under the G2O graph-based nonlinear optimization framework, extended Kalman filter is introduced to jointly optimize the multi-source sensor data to obtain the attitude and position estimation results of the UAV. Step 4: Determine whether the total length of the initial inspection path is greater than the single-use endurance of the UAV. If so, divide the UAV inspection task into segments according to the preset distance and return to Step 2 to replan the path; if not, output the initial inspection path as the final inspection path.

2. The method according to claim 1, characterized in that, Step 1, establishing a vehicle-machine collaborative inspection path planning model, selects the minimum set of parking points while ensuring coverage of all poles to be inspected, and aims to minimize the total travel time of the inspection vehicle. Under multiple objective constraints, it generates the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle. Specifically, this includes: Obtain the set of maintenance stations, alternative parking points, and towers to be inspected within the inspection area; Let the optimization objective be expressed as: ； Among them, Z total Z represents the total operating time of the inspection vehicle. q Let q be the working time of the inspection vehicle on the qth working day, where q∈W and W is the set of working days; Suppose that the multi-objective constraints include the working time constraint, expressed as: ； ; Where P is the set of parking points, h i The value is 1 if the i-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise. i Let h be the parking time of the inspection vehicle at parking point i, i.e., the inspection time at parking point i, where i∈P, G is the set of vehicle waypoints, G=P∪A, and A is the set of maintenance stations. i’ Indicates the i-th ′ The value is 1 when an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. j’ Indicates the j-th ′ The value is 1 if an alternative parking spot or maintenance station is selected as a vehicle waypoint, and 0 otherwise. qi’j’ For the inspection vehicle on the qth working day, from vehicle route point i ′ Driving to vehicle waypoint j ′ The value is 1 if it is true, and 0 otherwise. i’j’ For vehicle path point i ′ and vehicle waypoint j ′ The geographical straight-line distance between them, i ′ ∈G,j ′ ∈G, V is the average speed of the inspection vehicle, Z max The maximum working hours per day for inspection personnel of power companies; Let the multi-objective constraint include the inspection vehicle leaving the maintenance station every workday and eventually returning to the maintenance station, expressed as: ； in, The value is 1 if the inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle travels from vehicle waypoint j to vehicle waypoint i on the qth working day, and 0 otherwise. Suppose that the multi-objective constraints include that each selected parking spot can only be visited once by the inspection vehicle on a given workday, and that the vehicle must leave the selected parking spot after visiting it, expressed as: ； ； in, The value is 1 if the j-th alternative parking spot is selected as the vehicle waypoint, and 0 otherwise. For the inspection vehicle on the qth working day, from vehicle route point i ’ The value is 1 when the vehicle reaches path point j, and 0 otherwise. The inspection vehicle travels from vehicle waypoint i to vehicle waypoint j on the qth working day. ’ The value is 1 if it is true, and 0 otherwise. Let the multi-objective constraint include that all poles to be inspected are served by one and only one selected parking point, expressed as: ； Where, r ti The value is 1 if the t-th pole to be inspected is served by the i-th alternative parking point, and 0 otherwise. T is the set of poles to be inspected. Suppose that the multi-objective constraint includes the two endpoints of any segment of a vehicle path on a certain workday, which must be visited by the inspection vehicle on that workday, expressed as: ； in, The value is 1 if the inspection vehicle visits vehicle path point i' on the qth working day, and 0 otherwise. The value is 1 when the inspection vehicle visits vehicle path point j' on the qth working day, and 0 otherwise. Suppose that the multi-objective constraint includes the condition that if a parking spot is visited by an inspection vehicle on a certain workday, then that parking spot must be selected as a parking spot, expressed as: ； Where, x qi The value is 1 if the inspection vehicle visits the i-th alternative parking point on the q-th working day, and 0 otherwise. Suppose that the multi-objective constraints include decision variable constraints: ； ； Calculate the parking time C of the inspection vehicle at parking point i. i : ； Calculate the mission time z of the drone k at parking point i. ik : ； ； ； Among them, s tk The value d indicates that the t-th pole to be inspected is 1 when it is inspected by drone k, and 0 otherwise. ti Let c be the straight-line distance between pole t and parking point i. t Let U be the inspection time of the drone at point t on the pole, and U be the drone aggregation time. Suppose that the multi-objective constraint also includes that each tower is visited by one and only one drone, and that at each selected parking point, the total number of drone takeoffs equals the total number of towers served by that parking point, expressed as: ； ； Among them, f ik Let $k$ be the number of times drone $k takes off from parking point $i$. Solve the vehicle-machine collaborative inspection path planning model to generate the vehicle path and the total operation time of the inspection vehicle.

3. The method according to claim 2, characterized in that, Step 2, for each selected parking point, generates the initial inspection path for the drone based on map data containing geospatial information and using the A* algorithm for initial path planning. This specifically includes: The information on distributed photovoltaic power station inspection objects is structured, and the set of distributed photovoltaic power station objects to be inspected is defined as P. all The distributed photovoltaic power stations to be inspected are respectively denoted as P1, P2, ..., P... n Let P be the elevation of the highest point where the nth distributed photovoltaic power station module to be inspected is installed. n h; Local target location marking is performed on the distributed photovoltaic power station objects to be inspected, and the points will be marked in the set P. all One or more distributed photovoltaic power stations to be inspected that are within a straight-line distance of less than a set value X are classified as a local target, denoted as A. n And measure the latitude and longitude information of the outer edge feature points of the local target; According to the set P all The straight-line distance between any two distributed photovoltaic power station objects to be inspected is used to determine the range of values ​​for a set value X. The minimum value of the set value X is the minimum straight-line distance between any two objects in the set, and the maximum value is the maximum straight-line distance between any two objects. Determine whether the outer edge feature point of the local target satisfies the following conditions: it belongs to the outer edge of the photovoltaic module area, the outer edge angle is not 180°, and there exists at least one straight line passing through the outer edge feature point that does not intersect with the photovoltaic module in the local target. Among the outer edge feature points that meet the conditions, the line connecting the two outer edge feature points that are farthest apart is selected as a diagonal of the local target rectangular inspection area, denoted as L1. Another diagonal L2, of the same length as diagonal L1 and coinciding with its center point, is constructed, and the angle between the two diagonals is adjusted to form a rectangular inspection area containing all the outer edge feature points within the local target, denoted as S. n ; Based on the size and focal length of the charge-coupled device (CCD) of the camera mounted on the UAV and the working height L of the UAV, the inspection angle of the UAV is determined, and the imaging area of ​​a single shot is determined by the angle. The starting point for the UAV inspection of the nth local target is determined as S. n s, The inspection endpoint is S n e and the elevation S of the highest point where the components within the local target are installed. n h, the S n h is the maximum value of the installation elevation of each distributed photovoltaic power station component to be inspected within the local target area; The local target rectangular inspection area S n The vertex is taken as the inspection starting point, and the long side of the rectangle connected to the inspection starting point is taken as the initial inspection route. The S-shaped inspection method is used to plan the path of the rectangular inspection area. With the imaging area stitching covering the local target as a constraint, the latitude and longitude information of the UAV inspection path is calculated to generate the initial inspection path of the UAV within the local target. After completing the initial inspection path planning for each local target, the straight-line distance between different local targets is calculated based on the latitude and longitude information of each local target. The Traveling Salesman Problem (TSP) is used to determine the shortest inspection path for the UAV to traverse each local target and return to the starting local target, thus forming the initial inspection path between multiple local targets.

4. The method according to claim 3, characterized in that, Step 2, which combines visual SLAM to achieve environmental mapping and real-time localization, specifically includes: During the flight of the UAV along the initial inspection path, continuous environmental image data is collected using an airborne visual sensor. Feature extraction and feature matching are performed on the environmental image data, and the relative motion information of the UAV during flight is estimated based on the feature correlation between adjacent image frames; Based on the relative motion information, the current pose of the UAV is updated in real time, and the UAV pose is fused with environmental feature information to construct a local map of the UAV flight environment. During the continuous flight of the UAV, the local map is updated and optimized in real time, and the real-time pose information of the UAV is output.

5. The method according to claim 4, characterized in that, Step 3, during the UAV's flight along the initial inspection path, utilizes the G2O graph optimization framework to integrate multi-source sensor estimation of the UAV's attitude and position, and combines it with an adaptive error compensation algorithm to achieve robust flight trajectory control under disturbed environments, ensuring operational stability and accuracy. Specifically, this includes: The attitude and position estimation of a UAV is based on the G2O graph optimization framework, which integrates multi-source sensor data. This includes: extracting feature points from the image data acquired by the UAV using the FAST feature point detection algorithm to achieve visual positioning of the UAV; combining acceleration and angular velocity data acquired by MEMS inertial devices, latitude and longitude positioning information provided by GPS, and altitude positioning information acquired by barometers, and introducing extended Kalman filtering under the G2O graph-based nonlinear optimization framework to jointly optimize the multi-source sensor data and obtain the attitude and position estimation results of the UAV. Based on the attitude and position estimation results, an adaptive law is used for robust attitude angle tracking control of the UAV flight trajectory. Specifically, for the time delay function... and According to the Lyapunove functional, the stable solution of the flight trajectory robust tracking control satisfies: ； Among them, h1 and h2 are normal values; By introducing a lateral rotation and pitch oscillation vector, the reference motion equation for UAV flight trajectory tracking is controlled in a closed loop, and the asymptotic stability of UAV flight trajectory tracking control is achieved based on the Lyapunove stability principle. Based on the adaptive fusion of multi-sensor information, the flight trajectory tracking error is compensated in real time, realizing flight trajectory tracking control and error correction compensation of UAV under uncertain airflow disturbance.

6. A vehicle-machine collaborative inspection vehicle path planning device based on multi-objective optimization, characterized in that, include: The generation module is used to establish a vehicle-machine collaborative inspection path planning model. It selects the minimum set of parking points with the premise of covering all poles to be inspected, and takes minimizing the total driving time of the inspection vehicle as the optimization objective. Under multiple objective constraints, it generates the vehicle path, the set of selected parking points, and the total operation time of the inspection vehicle. The path planning module is used to generate the initial inspection path of the UAV for each selected parking point based on map data containing geospatial information and using the A-Star algorithm. It combines visual SLAM to realize environmental mapping and real-time positioning, and uses the improved A-Star algorithm and scrolling window mechanism for dynamic obstacle avoidance. The dynamic obstacle avoidance using the improved A* algorithm and the scrolling window mechanism specifically includes: During the flight of the UAV along the initial inspection path, the local environment map constructed based on visual SLAM and the real-time pose information of the UAV are used to obtain information on environmental changes ahead of the UAV's flight path in real time. Based on the environmental change information, an environmental risk cost reflecting obstacle distribution and safe distance constraints is introduced into the path cost function of the A-Star algorithm to dynamically evaluate the local cost of the UAV flight path and form an improved path search model. Starting from the current pose of the drone, the drone's flight path is locally replanned within the local space covered by the preset scrolling window to generate a local flight path that avoids dynamic obstacles. As the pose information of the UAV is updated in real time during flight, the position of the scrolling window is dynamically updated, and the local path replanning process is repeatedly executed. In this way, without changing the overall inspection sequence, continuous dynamic obstacle avoidance of the UAV during the inspection process is achieved, so as to output a safe and feasible inspection flight path. The control module is used to estimate the UAV's attitude and position based on the G2O graph optimization framework and multi-source sensor data during the UAV's flight along the initial inspection path. Combined with an adaptive error compensation algorithm, this achieves robust flight trajectory control under disturbed environments, ensuring operational stability and accuracy. The estimation of the UAV's attitude and position based on the G2O graph optimization framework and multi-source sensor data includes: The FAST feature point detection algorithm is used to extract feature points from the image data acquired by the UAV to achieve visual positioning of the UAV. The acceleration and angular velocity data acquired by MEMS inertial devices, the latitude and longitude positioning information provided by GPS, and the altitude positioning information acquired by barometer are combined. Under the G2O graph-based nonlinear optimization framework, extended Kalman filter is introduced to jointly optimize the multi-source sensor data to obtain the attitude and position estimation results of the UAV. The judgment output module is used to determine whether the total length of the initial inspection path is greater than the single-use endurance of the UAV. If so, the UAV inspection task is segmented according to a preset distance and returned to the path planning module to replan the path; if not, the initial inspection path is output as the final inspection path.

7. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 5.

Images