Coverage path planning method and device for heterogeneous unmanned aerial vehicle, equipment and medium

Abstract

The invention discloses a coverage path planning method and device for heterogeneous unmanned aerial vehicles, equipment and a medium. The method comprises the following steps: acquiring motion data corresponding to a plurality of heterogeneous unmanned aerial vehicles; performing performance evaluation on each heterogeneous unmanned aerial vehicle according to the motion data, and generating an unmanned aerial vehicle performance evaluation vector; performing region division on a to-be-planned region according to the unmanned aerial vehicle performance evaluation vector to obtain a plurality of task regions, respectively constructing a minimum spanning tree, and respectively generating an initial task coverage path according to the structure of each minimum spanning tree; and when an unmanned aerial vehicle task sequence is received, iteratively optimizing each initial task coverage path by adopting a simulated annealing algorithm according to the task allocation target function and the unmanned aerial vehicle performance evaluation vector, and generating a target task coverage path. Therefore, the use scene of the heterogeneous unmanned aerial vehicle coverage path is expanded, and the flexibility of unmanned aerial vehicle cooperative coverage path planning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and in particular to a coverage path planning method, device, equipment and medium for heterogeneous UAVs. Background Art

[0002] In typhoon disaster scenarios, intelligent patrol systems face challenges such as insufficient automation and low efficiency. To prevent typhoon disasters in complex power grid scenarios, rapidly respond to disasters, and accurately estimate losses, comprehensive mapping of typhoon-affected areas is necessary. This approach creates an accurate data foundation for emergency disaster assessments, facilitating monitoring, diagnosis, and assessment of power facilities after typhoon disasters. This, in turn, builds an intelligent and efficient power grid typhoon emergency disaster assessment system, effectively minimizing grid losses.

[0003] To this end, multi-UAV cooperative flight mission planning technology can be used to construct a multi-UAV collaborative inspection route planning strategy. Multi-UAV Cooperative Coverage Path Planning (MUCPP) is a key technology for improving the efficiency of UAVs in performing reconnaissance, surveillance, and search missions. By leveraging the parallelism and fault tolerance of multiple UAVs, it can effectively improve the speed and quality of inspection route completion. Currently, commonly used multi-UAV coverage path planning algorithms mainly include decision optimization algorithms such as market and auction algorithms, but rarely consider the performance differences between individual UAVs and sensors.

[0004] However, the coverage path area division and allocation algorithms proposed in the existing technology are mainly targeted at rectangular areas or are only applicable to the area to be covered and the area where the obstacle boundary is parallel to the rectangular side. They only focus on the path optimization of a single UAV and ignore the possibility of multi-UAV collaboration, resulting in limitations in usage scenarios and inability to flexibly plan UAV collaborative coverage paths. Summary of the Invention

[0005] The present invention provides a coverage path planning method, device, equipment and medium for heterogeneous drones, which solves the technical problem that existing drone planning algorithms are only applicable to rectangular areas or areas where the boundaries of the area to be covered and the obstacle are parallel to the rectangular edges, have limitations in usage scenarios, and cannot flexibly plan drone collaborative coverage paths.

[0006] A first aspect of the present invention provides a coverage path planning method for heterogeneous UAVs, comprising:

[0007] Obtain motion data corresponding to multiple heterogeneous drones;

[0008] Performing performance evaluation on each of the heterogeneous UAVs according to the motion data to generate a UAV performance evaluation vector;

[0009] Divide the planned area into regions according to the UAV performance evaluation vector to obtain multiple task areas and construct minimum spanning trees for each of the regions, and generate initial task coverage paths according to the structures of the minimum spanning trees;

[0010] When a UAV task sequence is received, a simulated annealing algorithm is used to iteratively optimize each of the initial task coverage paths according to the task allocation objective function and the UAV performance evaluation vector to generate a target task coverage path.

[0011] Optionally, the UAV performance evaluation vector includes a multi-dimensional mission performance index, and the mission performance index is:

[0012]

[0013] in, is the mission performance index of the i-th heterogeneous UAV, is the cruising altitude of the i-th heterogeneous UAV, is the cruising speed of the i-th heterogeneous UAV, is the horizontal field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the vertical field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the installation angle of the imaging sensor carried by the i-th heterogeneous UAV, is the cruising pitch angle of the i-th heterogeneous UAV, and m is the total number of heterogeneous UAVs.

[0014] Optionally, dividing the planned area according to the UAV performance evaluation vector to obtain multiple task areas and constructing minimum spanning trees for each area, and generating initial task coverage paths according to the structure of each minimum spanning tree, includes:

[0015] If the difference between any two task performance indexes is less than a preset performance threshold, the planning area is divided into regions according to the first cost function to obtain multiple task regions;

[0016] If the difference between any two task performance indexes is not less than a preset performance threshold, the planning area is divided and iteratively adjusted according to the second cost function to obtain multiple task areas;

[0017] Using key locations within each task area as nodes and determining connection weights according to the distances between any two nodes;

[0018] Traversing the nodes, selecting the nodes with the smallest weights in turn for connection, and obtaining a minimum spanning tree;

[0019] A starting node is selected from the minimum spanning tree, and each of the nodes is connected along the structure of the minimum spanning tree to obtain an initial task coverage path.

[0020] Optionally, if the gap between any two task performance indices is not less than a preset performance threshold, the planning area is divided and iteratively adjusted according to the second cost function to obtain multiple task areas, including:

[0021] If the difference between any two of the task performance indices is not less than a preset performance threshold, the target area ratio is initialized according to the ratio between the task performance indices;

[0022] Dividing the area to be planned into multiple regional shares;

[0023] Iteratively adjusting the share of each region according to a preset adjustment gradient to reduce the gap between the share of each region and the share of the target region;

[0024] Calculating a cost function value according to a second cost function;

[0025] When the cost function value converges, the area to be planned is divided according to the area share at the current moment to obtain multiple task areas;

[0026] The second cost function is:

[0027]

[0028] in, is the cost function value, is the regional share of the i-th heterogeneous UAV, is the target area proportion of the i-th heterogeneous UAV, is the total number of heterogeneous drones.

[0029] Optionally, after traversing the nodes and sequentially selecting nodes with the smallest weights for connection to obtain a minimum spanning tree, the method further includes:

[0030] The minimum spanning tree is updated according to a plurality of preset weight combinations to obtain a plurality of intermediate spanning trees;

[0031] Select the intermediate spanning tree with the minimum number of turns as the new minimum spanning tree.

[0032] Optionally, when receiving the UAV task sequence, the simulated annealing algorithm is used to iteratively optimize each of the initial task coverage paths according to the task allocation objective function and the UAV performance evaluation vector to generate a target task coverage path, including:

[0033] When receiving a UAV task sequence, determining whether the UAV performance evaluation vector meets the performance constraint condition;

[0034] If so, an initial decision variable matrix is created according to the UAV task sequence and each of the initial task coverage paths;

[0035] Calculating a first objective function value according to the initial decision variable matrix and the task allocation objective function;

[0036] Performing a perturbation operation on the initial decision variable matrix to obtain a neighborhood decision variable matrix;

[0037] Calculating a second objective function value according to the neighborhood decision variable matrix and the task allocation objective function;

[0038] Iteratively optimize the initial decision variable matrix according to the first objective function value and the second objective function value until the temperature drops to a termination temperature or the number of iterations reaches a maximum value, and use the initial decision variable matrix at the current moment as the target decision variable matrix;

[0039] The target decision variable matrix is parsed to obtain the target task coverage path corresponding to each of the heterogeneous UAVs.

[0040] Optionally, the task allocation objective function is:

[0041]

[0042] in, is the objective function value, is the time cost of heterogeneous drones, is the total flight cost of heterogeneous UAVs, To balance the workload, Regional rewards for heterogeneous drones, 、 、 and are weight parameters respectively.

[0043] A second aspect of the present invention provides a coverage path planning device for heterogeneous UAVs, comprising:

[0044] A data acquisition module is used to obtain motion data corresponding to multiple heterogeneous drones;

[0045] a performance evaluation module, configured to perform performance evaluation on each of the heterogeneous UAVs according to the motion data and generate a UAV performance evaluation vector;

[0046] A path generation module is used to divide the planning area into regions according to the UAV performance evaluation vector, obtain multiple task areas and construct minimum spanning trees for each of the regions, and generate initial task coverage paths according to the structure of each minimum spanning tree;

[0047] The path optimization module is used to iteratively optimize each of the initial task coverage paths using a simulated annealing algorithm according to the task allocation objective function and the drone performance evaluation vector when receiving the drone task sequence, so as to generate a target task coverage path.

[0048] A third aspect of the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the coverage path planning method for a heterogeneous UAV as described in any one of the first aspects of the present invention.

[0049] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the coverage path planning method for heterogeneous UAVs as described in any one of the first aspects of the present invention.

[0050] It can be seen from the above technical solutions that the present invention has the following advantages:

[0051] The present invention obtains motion data corresponding to multiple heterogeneous drones; performs performance evaluation on each heterogeneous drone based on the motion data to generate a drone performance evaluation vector; then divides the planned area into multiple task areas based on the drone performance evaluation vector to construct minimum spanning trees for each area; and generates initial task coverage paths based on the structure of each minimum spanning tree. Upon receiving a sequence of drone tasks, the simulated annealing algorithm is used to iteratively optimize each initial task coverage path based on the task allocation objective function and the drone performance evaluation vector to generate a target task coverage path. This expands the use cases of heterogeneous drone coverage paths and improves the flexibility of drone collaborative coverage path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A flowchart of the steps of a coverage path planning method for heterogeneous UAVs provided in an embodiment of the present invention;

[0054] Figure 2 A panoramic view of a heterogeneous drone detection area provided by an embodiment of the present invention;

[0055] Figure 3 A side view and a top view of a heterogeneous drone detection area provided by an embodiment of the present invention;

[0056] Figure 4 A schematic diagram of a rectangular area coverage process of a heterogeneous UAV provided in an embodiment of the present invention;

[0057] Figure 5 A schematic diagram of the camera field of view of an imaging sensor carried by a heterogeneous drone provided in an embodiment of the present invention;

[0058] Figure 6 A schematic diagram of the overlapping coverage areas of the imaging sensors carried by two heterogeneous drones provided in an embodiment of the present invention in the field of view;

[0059] Figure 7 A schematic diagram of the iterative process of task area division provided by an embodiment of the present invention;

[0060] Figure 8 A schematic diagram of a region division result provided by an embodiment of the present invention;

[0061] Figure 9 Schematic diagrams of various minimum spanning trees provided by embodiments of the present invention;

[0062] Figure 10 This is a structural block diagram of a coverage path planning device for heterogeneous UAVs provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] Embodiments of the present invention provide a coverage path planning method, apparatus, device, and medium for heterogeneous drones, which are used to solve the technical problem that existing drone planning algorithms are only applicable to rectangular areas or areas where the boundaries of the area to be covered and the obstacle are parallel to the rectangular edges, have limitations in usage scenarios, and cannot flexibly plan drone collaborative coverage paths.

[0064] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0065] See also Figure 1 , Figure 1A flowchart of the steps of a coverage path planning method for heterogeneous UAVs provided in an embodiment of the present invention.

[0066] The present invention provides a coverage path planning method for heterogeneous UAVs, comprising:

[0067] Step 101: Obtain motion data corresponding to multiple heterogeneous drones;

[0068] Motion data refers to the motion properties of heterogeneous UAVs during navigation, including but not limited to heading angle, flight speed, minimum turning radius, pitch angle limit, camera resolution, etc.

[0069] In this embodiment, the motion data corresponding to multiple heterogeneous drones are obtained as the data basis for subsequent coverage path planning. Specifically, the following formulas represent the set of drones and cameras respectively.

[0070]

[0071] U represents A collection of drones, Is the index of the drone. Represents a set of camera parameters. Each drone is equipped with a camera with a corresponding number. Its motion data is shown in Table 1 below:

[0072] Table 1

[0073]

[0074] Step 102 , performing performance evaluation on each heterogeneous UAV according to the motion data to generate a UAV performance evaluation vector;

[0075] Optionally, the drone performance evaluation vector includes a multi-dimensional mission performance index, where the mission performance index is:

[0076]

[0077] in, is the mission performance index of the i-th heterogeneous UAV, is the cruising altitude of the i-th heterogeneous UAV, is the cruising speed of the i-th heterogeneous UAV, is the horizontal field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the vertical field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the installation angle of the imaging sensor carried by the i-th heterogeneous UAV, is the cruising pitch angle of the i-th heterogeneous UAV, and m is the total number of heterogeneous UAVs.

[0078] In this embodiment, the UAV performs an area coverage task in three-dimensional space, and its imaging sensor detects an area on the ground such as Figure 2-Figure 3 shown. Figure 2 Display drone panorama, Figure 3 (a) is a top view, Figure 3 (b) is a side view. The trapezoidal area ABCD is the detection area of the imaging sensor on the ground. The parameters in the figure are defined as follows:

[0079] h: flight altitude, v: flight speed, : horizontal field of view, : Vertical field of view, : Front scan width, : rear scan width, : detection area length, θ: pitch angle, : The installation angle of the imaging sensor, that is, the angle between the longitudinal axis of the body and the vertical field of view, : The angle between the angle bisectors, : The distance between the UAV's projection on the ground and the center of the detection area.

[0080] according to Figure 2-Figure 3 , the calculation formula of the scanning width after the detection area can be deduced as:

[0081]

[0082] In this embodiment, the rear scan width is selected as the scan width for the UAV coverage path planning to avoid blurring of the edges of the reconnaissance image.

[0083] like Figure 4 As shown, Figure 4 The diagram shows the process of covering a rectangular area. The x and y in the figure represent the horizontal and vertical axes in the earth's axis system respectively; the rectangle The coverage path planning process for a single UAV is to find an optimal route that traverses the entire coverage area after scanning the area with a scan width w2. Line segment L is the optimal coverage route planned within the area. The goal of multi-UAV collaborative path planning is to minimize the number of turns required by multiple UAVs while ensuring that each UAV obtains the optimal route.

[0084] Assume that the sweep width of the UAV Ui is ω 1,2 , the coverage area is polygon P, and its area is area. In order to ensure the imaging quality and resolution, the cruising speed of the UAV Ui is set to v 0,i , the cruising altitude is h 0,i , the pitch angle during cruising is θ 0,i ; The vertical field of view of the imaging sensor it carries is 1,i , the horizontal field of view is 1,i , the installation angle is 1,i The coverage task is completed simultaneously by multiple heterogeneous UAVs, and the time is t. According to the multi-UAV time coordination goal, the following equation can be established:

[0085]

[0086] Combining the above two equations, we can get:

[0087]

[0088] After the change, we can get:

[0089]

[0090] Then the mission performance index of UAV Ui is

[0091]

[0092] The performance evaluation vector of multiple UAVs performing coverage tasks is:

[0093]

[0094] It should be noted that P Ui represents the initial three-dimensional coordinates of Ui; φ i is the heading angle of Ui; Vi is the flight speed of Ui; Ri is the minimum turning radius of Ui. [ω li ,ω hi ] is the pitch angle limit of Ui, where ω li <0,ω hi >0.

[0095] E Ui It is the maximum flight time of Ui at a fixed flight speed. li ,H hi ] is the height range that Ui can reach to perform the task. The detection resolution range of Ui is defined as [P rli ,P rhi ], which can be determined by the properties of the sensor. When Ui flies to H li When its resolution is P rhi When Ui flies to h i When its resolution is P rli The upper limit of the number of targets covered by Ui is L i . Predefined parameter C Ui ∈[0,1] to represent the capability of Ui. Therefore, C Ui / C TjIt represents the success rate of Ui detecting the UAV task Tj.

[0096] In this embodiment, for the imaging sensor carried by the heterogeneous UAV, if the sub-region is completely within the field of view of the camera, the mapping task of the region is completed. The field of view of the camera is assumed to be the rectangular area below the UAV, such as Figure 5 The actual effect of Ui on Tj is affected by its parameters. These variables are denoted by the subscript ij.

[0097] H ij represents the actual flight height from Ui to Tj, where [v ij , w ij ] represents the front and side widths of the real field of view when Ui is at Tj. When Ui is at Tj, the overlap ratio of the two photos in the field of view is the front overlap ratio OL ij ,like Figure 6 Similarly, the overlap ratio between the two paths is the side overlap ratio SL ij .

[0098]

[0099] Respectively represent the number of pixels in front and on the sides of the camera. is the focal length of the camera (unit: mm); is the unit pixel size (unit: um); and is the shutter frame rate (unit: s).

[0100] Step 103: Divide the planned area into regions according to the UAV performance evaluation vector to obtain multiple mission regions and construct minimum spanning trees for each region. Then, generate initial mission coverage paths according to the structures of each minimum spanning tree.

[0101] In this example, offline planning of coverage mapping paths for a multi-UAV system is performed to address the problem of strongly coupled constraints. This approach is solved by decomposing the mCPP problem into multiple single-robot CPP problems. This iterative process aims to construct an assignment matrix in which all cells are labeled as corresponding to a unique UAV. After the area is divided, each UAV is responsible for covering a specific portion of the grid. This assignment of dedicated sub-areas ensures the generation of disjoint paths, allowing multiple UAVs to safely operate simultaneously and avoid potential collisions.

[0102] In an example of the present application, step 103 may include the following steps S11-S15:

[0103] S11. If the difference between any two task performance indices is less than a preset performance threshold, the planning area is divided into regions according to the first cost function to obtain multiple task regions;

[0104] In this embodiment, if the gap between any two mission performance indices is less than a preset performance threshold, the planned area is divided into regions according to the first cost function to obtain multiple mission regions. Through fair region division, all drones will be assigned to cover the same percentage of the area.

[0105] Among them, the first cost function is:

[0106]

[0107]

[0108] in, and f represent the current share and global fair share of the i-th drone, respectively. Unoccupied cells.

[0109] S12. If the difference between the performance indexes of any two tasks is not less than a preset performance threshold, the planning area is divided and iteratively adjusted according to the second cost function to obtain multiple task areas.

[0110] like Figure 7 As shown, Figure 7 The allocation of four drones to one area during the execution time is Figure 7 The initial position shown in (a) is obtained by Figure 7 (b)- Figure 7 (f) is continuously iterated and adjusted to obtain multiple task areas such as Figure 7 (f) Figure 7 (b)- Figure 7 (e) Schematic diagram of the task area with different progress during the iteration process. The original version provides a fair area division, which means that all drones will be assigned to cover the same percentage of the area. However, when the gap between the performance index of any two tasks is not less than the preset performance threshold, the percentage is user-defined to facilitate the simultaneous operation of heterogeneous drones in the same task. These drones have different energy and operational capabilities. The area division results are shown in the figure below. Figure 8 As shown in the figure, an example of path generation is shown for proportional area allocation of the area to be planned based on a user-defined percentage, where the area to be planned is divided into three parts of 20%, 30% and 50%, which are respectively allocated to different heterogeneous UAVs for generating the initial mission coverage path.

[0111] In an example of the present application, S12 may include the following sub-steps:

[0112] If the gap between any two task performance indices is not less than the preset performance threshold, the target area ratio is initialized according to the ratio between the task performance indices;

[0113] Divide the area to be planned into multiple regional shares;

[0114] Iteratively adjust the share of each region according to the preset adjustment gradient to reduce the gap between the share of each region and the share of the target region;

[0115] Calculating a cost function value according to a second cost function;

[0116] When the cost function value converges, the area to be planned is divided according to the regional share at the current moment to obtain multiple task areas;

[0117] The second cost function is:

[0118]

[0119] in, is the cost function value, is the regional share of the i-th heterogeneous UAV, is the target area proportion of the i-th heterogeneous UAV, is the total number of heterogeneous drones.

[0120] S13, using key locations in each task area as nodes and determining connection weights according to the distances between each node;

[0121] S14, traverse the nodes, select the nodes with the smallest weights in turn for connection, and obtain the minimum spanning tree;

[0122] S15. Select a starting node from the minimum spanning tree, connect each node along the structure of the minimum spanning tree, and obtain an initial task coverage path.

[0123] Once each robot is assigned to its dedicated portion of the overall area, a single-drone problem is solved for each mission area. For path generation, for each drone's dedicated sub-area, key locations within the mission area, such as monitoring targets (e.g., power facilities and disaster sites), are treated as nodes. The distances between nodes (Euclidean distances can be calculated using geographic coordinates) serve as edge weights to construct a graph structure. Based on the minimum spanning tree structure constructed for each mission area, the initial mission coverage path for the corresponding drone is generated. A minimum spanning tree (MST) is constructed for each mission area, and a path is then generated around the tree.

[0124] Furthermore, after traversing the nodes and sequentially selecting the nodes with the smallest weights for connection to obtain the minimum spanning tree, step 103 further includes:

[0125] The minimum spanning tree is updated according to a plurality of preset weight combinations to obtain multiple intermediate spanning trees;

[0126] Select the intermediate spanning tree with the minimum number of turns as the new minimum spanning tree.

[0127] Due to the spanning tree nature of the algorithm, the path is likely to contain many unnecessary turns if no measures are taken to avoid them. The number of turns is one of the main factors affecting flight time and energy consumption. To solve this problem, a set of n different weight combinations are tested for each mission area to control the connection of nodes in the spanning tree, and the minimum spanning tree setting that produces the path with the fewest turns is selected from them. In this work, four different combinations are tested for each sub-area, corresponding to the spanning trees connected at the upper, lower, rightmost and leftmost levels available, such as Figure 9 shown.

[0128] Step 104: When the UAV task sequence is received, the simulated annealing algorithm is used to iteratively optimize each initial task coverage path according to the task allocation objective function and the UAV performance evaluation vector to generate a target task coverage path.

[0129] In an embodiment of the present application, a drone task sequence generated based on specific mission requirements is received. For example, in a city's power inspection mission, the task sequence may include information about each substation and transmission line node that needs to be inspected. These task points are sorted according to certain rules (such as geographic location and importance), forming a task sequence that the drone must visit in sequence. The performance of each drone is evaluated to obtain a performance evaluation vector. After iterative optimization using a simulated annealing algorithm, the final solution obtained is the target task coverage path. This path comprehensively considers the task allocation objective function and the drone performance evaluation vector, achieving a good balance between time cost, total flight cost, balanced workload, and regional value rewards, thereby improving the efficiency and quality of drone mission execution.

[0130] In one example of the present application, step 104 may include the following sub-steps:

[0131] When receiving the UAV task sequence, it is determined whether the UAV performance evaluation vector meets the performance constraint conditions;

[0132] If so, create an initial decision variable matrix according to the UAV task sequence and the coverage path of each initial task;

[0133] Calculate the first objective function value according to the initial decision variable matrix and the task allocation objective function;

[0134] Perform a perturbation operation on the initial decision variable matrix to obtain the neighborhood decision variable matrix;

[0135] Calculate the second objective function value according to the neighborhood decision variable matrix and the task allocation objective function;

[0136] According to the first objective function value and the second objective function value, the initial decision variable matrix is iteratively optimized until the temperature drops to the termination temperature or the number of iterations reaches the maximum value, and the initial decision variable matrix at the current moment is used as the target decision variable matrix;

[0137] Analyze the target decision variable matrix to obtain the target mission coverage path corresponding to each heterogeneous UAV.

[0138] Optionally, the task allocation objective function is:

[0139]

[0140] in, is the objective function value, is the time cost of heterogeneous drones, is the total flight cost of heterogeneous UAVs, To balance the workload, Regional rewards for heterogeneous drones, 、 、 and are weight parameters respectively.

[0141] In this embodiment of the present application, when a drone task sequence is received, it is determined whether the drone performance evaluation vector meets the performance constraint conditions; if so, an initial decision variable matrix X is created according to the drone task sequence and each initial task coverage path, specifically:

[0142] The multi-heterogeneous UAV collaborative mapping problem is extended to CMTAP, where the goal is to find the best combination of decision variables. Therefore, a binary decision variable matrix is defined as:

[0143]

[0144] =1 means U i Existence from T j to T k Otherwise, =0. Binary decision variables are directional; when (j, k) = 1, (k, j) = 0. Performance constraints may include, but are not limited to, mileage constraints, flight altitude constraints, drone capability constraints, mission requirement constraints, and allocation constraints. Specifically:

[0145] Mileage constraints:

[0146] Ui The total mileage cannot exceed the maximum endurance time.

[0147]

[0148] is the distance the drone flies, is the speed of the drone, The maximum flight time of the drone,

[0149] Flight altitude constraints:

[0150] Different drones have their own flight altitude ranges, and mapping tasks must be completed within these altitude ranges.

[0151]

[0152] is the minimum flight altitude of the i-th UAV, is the maximum flight altitude of the i-th UAV, is the flight altitude of the i-th UAV in the j-th mission.

[0153] Drone capability constraints:

[0154] The mission success rate can be calculated by dividing the UAV capability parameter by the mission difficulty coefficient, and the mission success rate must be greater than the predefined parameter.

[0155]

[0156] Task requirements and constraints:

[0157] The detection performance of the UAV should meet the requirements of resolution, lateral overlap rate and forward overlap rate.

[0158]

[0159] in, is the forward overlap rate of the i-th UAV in the j-th mission, is the forward overlap rate required for the jth task, is the lateral overlap rate of the i-th UAV in the j-th mission, is the lateral overlap rate required for the jth task, is the drone resolution, Required resolution for the task

[0160] Allocation constraints:

[0161] For the CMTAP model, the binary decision variables The task allocation constraints are defined as follows:

[0162]

[0163] is the total number of tasks, is the total number of heterogeneous drones.

[0164] In this example, a simulated annealing algorithm is used to solve the global path optimization problem for multi-UAV coordinated coverage path planning. This algorithm simulates the temperature changes in a physical annealing process, gradually reducing the "energy" of the solution to find the optimal path allocation solution. The specific application steps are as follows:

[0165] (1) Initialize parameters and generate initial solution. Set the initial temperature for the algorithm , temperature update rate α, termination temperature and the maximum number of iterations Based on the current mission area and the performance characteristics of the UAVs, an initial solution x is randomly generated, representing the mission allocation and path planning scheme for each UAV. The objective function value f(x) of the initial solution is calculated as the target value of the current solution x and saved as the current optimal solution.

[0166] The objective function is the weighted sum of the sub-objectives of step S41 above:

[0167]

[0168] (2) Generate and evaluate neighborhood solutions. In each iteration, randomly select a neighborhood solution x' of the current solution x. Specifically, swap the allocation of any two sub-areas in the current solution, that is, swap the mission areas of two drones, or perform a small-scale perturbation operation on the coverage path sequence of a certain drone. The new neighborhood solution x' represents a different task allocation or path planning scheme. Calculate the objective function value f(x') of the neighborhood solution and compare it with the objective value f(x) of the current solution.

[0169] (3) The Metropolis criterion is used to decide whether to accept the new solution. When the objective function value f(x') of the neighborhood solution x' is less than the current solution f(x), it means that the new task allocation scheme is better than the current solution, so the new solution x' is directly accepted as the current solution; if f(x') is greater than f(x), that is, the neighborhood solution is worse than the current solution, then according to the Metropolis criterion, the inferior solution is accepted with probability P. The probability of accepting the inferior solution gradually decreases as the temperature decreases. This allows for greater exploration in the early stages to prevent falling into local optimality, while focusing on local search in the later stages for precise optimization.

[0170] The acceptance criterion (Metropolis criterion) P is defined as:

[0171]

[0172] Represents the difference in scalar function values. is the objective function value of the new solution (neighborhood solution). is the objective function value of the current solution. , it means that the new solution is worse than the current solution, that is, it is an inferior solution.

[0173] (4) Update temperature. After each iteration, the algorithm uses geometric cooling to update the temperature:

[0174]

[0175] Where α is the cooling rate (ranging from 0.8 to 0.99). By gradually lowering the temperature, the algorithm gradually transitions from a global search to a local search, making the solution more stable and moving toward the global optimal solution.

[0176] (5) Check the termination conditions and output the optimal solution. When the temperature drops to the termination temperature Below, or reaching the maximum number of iterations The algorithm stops when . At this point, the current solution x is considered to be the global near-optimal solution and is returned as the final task allocation and path planning solution. The ultimate goal is to minimize the total path length, balance the task loads of each UAV, and minimize the number of turns and total flight time, thereby achieving efficient coverage and collaborative operations of the UAV system.

[0177] Through these steps, the simulated annealing algorithm achieves dynamic global optimization in multi-UAV coverage path planning. The algorithm's objective function considers multiple optimization goals and comprehensively weighs the pros and cons of solutions at each iteration, gradually approaching the global optimal solution.

[0178] In an embodiment of the present application, motion data corresponding to multiple heterogeneous drones is acquired; performance evaluation of each heterogeneous drone is performed based on the motion data to generate a drone performance evaluation vector; the planned area is divided according to the drone performance evaluation vector to obtain multiple task areas, and minimum spanning trees are constructed for each area. Initial task coverage paths are generated based on the structure of each minimum spanning tree; when a drone task sequence is received, a simulated annealing algorithm is used to iteratively optimize each initial task coverage path based on the task allocation objective function and the drone performance evaluation vector to generate a target task coverage path. This expands the use cases of heterogeneous drone coverage paths and improves the flexibility of drone collaborative coverage path planning. This approach not only considers path optimization for a single drone, but also coordinates multiple drones to achieve global optimal coverage, ensuring mission completion in the shortest possible time and with the highest efficiency. By optimizing coverage paths and coordination between drones, overall mission execution time and operating costs are reduced. Through effective path planning, drones can reduce flight distance and flight time, thereby saving energy and maintenance costs.

[0179] like Figure 10 As shown, Figure 10 This is a structural block diagram of a coverage path planning device for heterogeneous UAVs provided in an embodiment of the present invention.

[0180] An embodiment of the present invention provides a coverage path planning device for heterogeneous UAVs, comprising:

[0181] The data acquisition module 901 is used to acquire motion data corresponding to multiple heterogeneous drones;

[0182] The performance evaluation module 902 is used to perform performance evaluation on each heterogeneous UAV according to the motion data and generate a UAV performance evaluation vector;

[0183] A path generation module 903 is configured to divide the planned area into regions according to the UAV performance evaluation vector, obtain multiple mission regions, construct minimum spanning trees for each region, and generate initial mission coverage paths according to the structure of each minimum spanning tree.

[0184] The path optimization module 904 is used to iteratively optimize each initial task coverage path using a simulated annealing algorithm according to the task allocation objective function and the UAV performance evaluation vector when receiving the UAV task sequence, and generate a target task coverage path.

[0185] Optionally, the drone performance evaluation vector includes a multi-dimensional mission performance index, where the mission performance index is:

[0186]

[0187] in, is the mission performance index of the i-th heterogeneous UAV, is the cruising altitude of the i-th heterogeneous UAV, is the cruising speed of the i-th heterogeneous UAV, is the horizontal field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the vertical field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the installation angle of the imaging sensor carried by the i-th heterogeneous UAV, is the cruising pitch angle of the i-th heterogeneous UAV, and m is the total number of heterogeneous UAVs.

[0188] Optionally, the path generation module 903 includes:

[0189] A first region division submodule is configured to divide the planned region into multiple task regions according to a first cost function if the difference between any two task performance indices is less than a preset performance threshold;

[0190] A second region division submodule is configured to divide the planned region into regions according to a second cost function and iteratively adjust the regions to obtain multiple task regions if the difference between the performance indexes of any two tasks is not less than a preset performance threshold;

[0191] The node connection submodule is used to use key locations in each task area as nodes and determine the connection weight according to the distance between each node;

[0192] The minimum spanning tree generation submodule is used to traverse the nodes and select the nodes with the smallest weight to connect in turn to obtain the minimum spanning tree;

[0193] The path planning submodule is used to select the starting node from the minimum spanning tree, connect each node along the structure of the minimum spanning tree, and obtain the initial task coverage path.

[0194] Optionally, the second area division submodule is specifically configured to:

[0195] If the gap between any two task performance indices is not less than the preset performance threshold, the target area ratio is initialized according to the ratio between the task performance indices;

[0196] Divide the area to be planned into multiple regional shares;

[0197] Iteratively adjust the share of each region according to the preset adjustment gradient to reduce the gap between the share of each region and the share of the target region;

[0198] Calculating a cost function value according to a second cost function;

[0199] When the cost function value converges, the area to be planned is divided according to the regional share at the current moment to obtain multiple task areas;

[0200] The second cost function is:

[0201]

[0202] in, is the cost function value, is the regional share of the i-th heterogeneous UAV, is the target area proportion of the i-th heterogeneous UAV, is the total number of heterogeneous drones.

[0203] Optionally, after traversing the nodes and sequentially selecting the nodes with the smallest weights for connection to obtain a minimum spanning tree, the path generation module 903 further includes a minimum spanning tree updating submodule, which is specifically used to:

[0204] The minimum spanning tree is updated according to a plurality of preset weight combinations to obtain multiple intermediate spanning trees;

[0205] Select the intermediate spanning tree with the minimum number of turns as the new minimum spanning tree.

[0206] Optionally, the path optimization module 904 is specifically configured to:

[0207] When receiving the UAV task sequence, it is determined whether the UAV performance evaluation vector meets the performance constraint conditions;

[0208] If so, create an initial decision variable matrix according to the UAV task sequence and the coverage path of each initial task;

[0209] Calculate the first objective function value according to the initial decision variable matrix and the task allocation objective function;

[0210] Perform a perturbation operation on the initial decision variable matrix to obtain the neighborhood decision variable matrix;

[0211] Calculate the second objective function value according to the neighborhood decision variable matrix and the task allocation objective function;

[0212] According to the first objective function value and the second objective function value, the initial decision variable matrix is iteratively optimized until the temperature drops to the termination temperature or the number of iterations reaches the maximum value, and the initial decision variable matrix at the current moment is used as the target decision variable matrix;

[0213] Analyze the target decision variable matrix to obtain the target mission coverage path corresponding to each heterogeneous UAV.

[0214] Optionally, the task allocation objective function is:

[0215]

[0216] in, is the objective function value, is the time cost of heterogeneous drones, is the total flight cost of heterogeneous UAVs, To balance the workload, Regional rewards for heterogeneous drones, 、 、 and are weight parameters respectively.

[0217] An embodiment of the present invention provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory. When the computer program is executed by the processor, the processor performs the steps of the coverage path planning method for a heterogeneous unmanned aerial vehicle as described in any embodiment of the present invention.

[0218] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, the coverage path planning method for heterogeneous UAVs as described in any embodiment of the present invention is implemented.

[0219] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0220] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0221] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of the present embodiment according to actual needs.

[0222] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or software functional modules.

[0223] If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing an electronic device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0224] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A coverage path planning method for heterogeneous UAVs, characterized in that: include: Obtain motion data corresponding to multiple heterogeneous drones; Performing performance evaluation on each of the heterogeneous UAVs according to the motion data to generate a UAV performance evaluation vector; Divide the planned area into regions according to the UAV performance evaluation vector to obtain multiple task areas and construct minimum spanning trees for each of the regions, and generate initial task coverage paths according to the structures of the minimum spanning trees; When a UAV task sequence is received, a simulated annealing algorithm is used to iteratively optimize each of the initial task coverage paths according to the task allocation objective function and the UAV performance evaluation vector to generate a target task coverage path.

2. The method according to claim 1, characterized in that The UAV performance evaluation vector includes a multi-dimensional mission performance index, which is: in, is the mission performance index of the i-th heterogeneous UAV, is the cruising altitude of the i-th heterogeneous UAV, is the cruising speed of the i-th heterogeneous UAV, is the horizontal field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the vertical field of view of the imaging sensor carried by the i-th heterogeneous UAV, is the installation angle of the imaging sensor carried by the i-th heterogeneous UAV, is the cruising pitch angle of the i-th heterogeneous UAV, and m is the total number of heterogeneous UAVs.

3. The method according to claim 2, characterized in that The method of dividing the planned area according to the UAV performance evaluation vector to obtain multiple task areas and constructing minimum spanning trees for each task area, and generating initial task coverage paths according to the structure of each minimum spanning tree, includes: If the difference between any two task performance indexes is less than a preset performance threshold, the planning area is divided into regions according to the first cost function to obtain multiple task regions; If the difference between any two task performance indexes is not less than a preset performance threshold, the planning area is divided and iteratively adjusted according to the second cost function to obtain multiple task areas; Using key locations within each task area as nodes and determining connection weights according to the distances between any two nodes; Traversing the nodes, selecting the nodes with the smallest weights in turn for connection, and obtaining a minimum spanning tree; A starting node is selected from the minimum spanning tree, and each of the nodes is connected along the structure of the minimum spanning tree to obtain an initial task coverage path.

4. The method according to claim 3, characterized in that If the gap between any two task performance indices is not less than a preset performance threshold, the planned area is divided and iteratively adjusted according to the second cost function to obtain multiple task areas, including: If the difference between any two of the task performance indices is not less than a preset performance threshold, the target area ratio is initialized according to the ratio between the task performance indices; Dividing the area to be planned into multiple regional shares; Iteratively adjusting the share of each region according to a preset adjustment gradient to reduce the gap between the share of each region and the share of the target region; Calculating a cost function value according to a second cost function; When the cost function value converges, the area to be planned is divided according to the area share at the current moment to obtain multiple task areas; The second cost function is: in, is the cost function value, is the regional share of the i-th heterogeneous UAV, is the target area proportion of the i-th heterogeneous UAV, is the total number of heterogeneous drones.

5. The method according to claim 3, characterized in that After traversing the nodes and sequentially selecting nodes with the smallest weights for connection to obtain a minimum spanning tree, the method further includes: The minimum spanning tree is updated according to a plurality of preset weight combinations to obtain a plurality of intermediate spanning trees; Select the intermediate spanning tree with the minimum number of turns as the new minimum spanning tree.

6. The method according to claim 1, characterized in that When receiving the UAV task sequence, the simulated annealing algorithm is used to iteratively optimize each of the initial task coverage paths according to the task allocation objective function and the UAV performance evaluation vector to generate a target task coverage path, including: When receiving a UAV task sequence, determining whether the UAV performance evaluation vector meets the performance constraint condition; If so, an initial decision variable matrix is created according to the UAV task sequence and each of the initial task coverage paths; Calculating a first objective function value according to the initial decision variable matrix and the task allocation objective function; Performing a perturbation operation on the initial decision variable matrix to obtain a neighborhood decision variable matrix; Calculating a second objective function value according to the neighborhood decision variable matrix and the task allocation objective function; Iteratively optimize the initial decision variable matrix according to the first objective function value and the second objective function value until the temperature drops to a termination temperature or the number of iterations reaches a maximum value, and use the initial decision variable matrix at the current moment as the target decision variable matrix; The target decision variable matrix is parsed to obtain the target task coverage path corresponding to each of the heterogeneous UAVs.

7. The method according to claim 6, characterized in that The task allocation objective function is: in, is the objective function value, is the time cost of heterogeneous drones, is the total flight cost of heterogeneous UAVs, To balance the workload, Regional rewards for heterogeneous drones, 、 、 and are weight parameters respectively.

8. A coverage path planning device for heterogeneous UAVs, characterized in that: include: A data acquisition module is used to obtain motion data corresponding to multiple heterogeneous drones; a performance evaluation module, configured to perform performance evaluation on each of the heterogeneous UAVs according to the motion data and generate a UAV performance evaluation vector; A path generation module is used to divide the planning area into regions according to the UAV performance evaluation vector, obtain multiple task areas and construct minimum spanning trees for each of the regions, and generate initial task coverage paths according to the structure of each minimum spanning tree; The path optimization module is used to iteratively optimize each of the initial task coverage paths using a simulated annealing algorithm according to the task allocation objective function and the drone performance evaluation vector when receiving the drone task sequence, so as to generate a target task coverage path.

9. An electronic device, characterized in that: It includes a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the coverage path planning method for a heterogeneous unmanned aerial vehicle as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the coverage path planning method for a heterogeneous UAV as described in any one of claims 1 to 7 is implemented.

Images