A Method for Multi-UAV Cooperative Area Coverage under Communication Constraints

Abstract

The present invention belongs to the technical field of UAV cooperative control and path planning, and particularly relates to a method for multi-UAV cooperative area coverage under communication constraints, including the following steps: map modeling and map updating; establishing the UAV motion equation; using the method of generating a dynamic window for path planning, setting the evaluation function threshold, and determining whether the maximum value of the evaluation function of each path is greater than the threshold. If it is greater, select the path with the highest evaluation function value and move forward according to the plan for a fixed duration; if it is less, track the unexplored area in the map to determine the target end point; S6. Use the improved A* algorithm to calculate the path of the UAV to the target end point, and combine the dynamic window algorithm to guide the UAV to execute the task. The present invention improves the coverage rate of multi-UAV cooperation and effectively reduces the coverage cost.

Description

Technical Field

[0001] The present invention belongs to the technical field of UAV cooperative control and path planning, and particularly relates to a method for multi-UAV cooperative area coverage under communication constraints. Background Art

[0002] With the rapid development of UAV technology, UAVs are increasingly widely used in emergency rescue. However, due to the relatively low performance of single UAVs, they cannot well complete tasks, and the area coordinated coverage of multi-UAVs has received wide attention. Multi-UAV cooperation can meet more needs and has the advantages of high efficiency, strong fault tolerance, and diversity. Summary of the Invention

[0003] The purpose of the present invention is to solve the above technical problems. Aiming at the problem of multi-UAV cooperative coverage under communication constraints, a method for multi-UAV cooperative area coverage under communication constraints is provided, which can achieve fast multi-UAV cooperative area coverage on the premise of meeting the dynamic characteristics of UAVs.

[0004] To solve the above technical problems, the technical solution adopted by the present invention is: A method for multi-UAV cooperative area coverage under communication constraints, comprising the following steps:

[0005] S1. Map modeling: Set the task area, create two maps and rasterize them at a fixed interval, and assign values to the grids according to different grid assignment methods to obtain an obstacle avoidance map and a coverage map respectively;

[0006] S2. Use the dynamic window algorithm for path planning, set the evaluation function threshold, and judge whether the maximum value of the evaluation function of each path is greater than the threshold. If it is greater, go to step S3; if it is less, go to step S4;

[0007] When using the dynamic window algorithm for path planning, the evaluation function adopted by each path is:

[0008] J = ω 1 J 1 + ω 2 J 2 + ω 3 J 3 ;

[0009] Wherein, J represents the evaluation function, J 1 is the normalized coverage return; J 2 is the normalized fuel consumption penalty; J 3 is the normalized repulsive force penalty between UAVs; ω 1 , ω 2 , ω 3 are the corresponding weights;

[0010] S3. Select the path with the highest evaluation function value, move forward for a fixed duration according to the plan, then update the map based on the position of the UAV, and return to step S2;

[0011] S4. Track the unexplored areas in the map, binarize the map according to the explored areas and other areas to obtain a binary matrix, track the connected areas of the binary matrix, obtain the list of boundary grids of the areas of each connected area, select the connected areas with an area greater than the area threshold, calculate the distance between the boundary grids and the UAV, and select the nearest grid as the target end point;

[0012] S5. Use the improved A* algorithm to calculate the path of the UAV to the target end point, and combine the dynamic window algorithm to guide the UAV to move forward for a fixed duration according to the plan until it reaches the end point of the path obtained by the improved A* algorithm, and then return to step S2; The improved A* algorithm includes the following steps:

[0013] Use the A* algorithm to find a complete route from the current position of the UAV to the target point, and obtain the list of nodes corresponding to this route;

[0014] Extract the key nodes from the list of nodes, and obtain a new list of nodes according to the key nodes;

[0015] Extract and remove the intermediate nodes from the new list of nodes to obtain the final list of nodes;

[0016] Take the final list of nodes as the target points of the dynamic window algorithm in turn, and guide the UAV to fly to the target end point to execute the task.

[0017] The specific method of extracting the key points from the list of nodes and obtaining a new list of nodes according to the key nodes is as follows:

[0018] Calculate the straight-line equation of the first node and the second node in the list of nodes;

[0019] Substitute the subsequent nodes in turn to judge whether they conform to the equation. If they conform, continue to judge the next node; if not, judge the previous node as the key node; generate a new straight-line equation according to the key node and the next node of the key node, and repeat the above steps until the end point;

[0020] Record the key nodes to form a new list of nodes.

[0021] The specific method of extracting the intermediate nodes from the new list of nodes is as follows:

[0022] Starting from obtaining the first node of the new node list, successively determine whether there are obstacles between it and non-adjacent nodes. If there are no obstacles, the nodes between these two nodes are intermediate nodes, and continue to determine whether there are obstacles between it and the next node; if there are obstacles, start from the previous node and continue the above steps until the end point.

[0023] In the step S1, the grid assignment formula of the obstacle avoidance map is:

[0024]

[0025] In the formula: Ω 1 is the set of prohibited areas; Ω 2 is the set of other areas; μ mn represents the value of the grid at the m-th row and n-th column of the obstacle avoidance map;

[0026] The grid assignment formula of the coverage map is:

[0027]

[0028] In the formula: Ω 3 (k) is the boundary area set of the undetected area; Ω 4 (k) is the set to be searched except the boundary area; Ω 5 (k) is the set of prohibited areas and the searched area, μ ij (k) represents the value of the grid at the i-th row and j-th column of the coverage map, and k represents the time.

[0029] In the step S3, the update method of the coverage map is:

[0030] (1) Obtain the first map map1 and the second map map2 respectively. The grid status of the first map map1 is:

[0031]

[0032] The grid status of the second map map2 is:

[0033]

[0034] In the formula: Ω 6 (k) is the area outside the detection range; Ω 7 (k) is the area within the detection range; Ω 8 (k) is the boundary set of the area within the detection range; Ω 9 (k) is the other area; μ mn represents the value of the grid at the m-th row and n-th column of the obstacle avoidance map;

[0035] (2) Obtain the coverage map with the UAV position as the center and a length of 2*Re The matrix map of -1; Re represents the radius of the detection range of the UAV;

[0036] (3) Multiply the corresponding elements of map and map1 to obtain the matrix mapt1;

[0037] (4) Multiply the corresponding elements of the new logicalized matrix of map and map2 to obtain the matrix mapt2;

[0038] (5) Add the matrix mapt1 and the matrix mapt2 to obtain a new map, and assign it to the corresponding position of the coverage map Map2.

[0039] In the step S3, the UAV updates its own map with the position information obtained during communication; the communication strategy of the UAV is:

[0040] Each UAV records each time node, the positions it has passed and the positions passed by other UAVs learned during communication, and establishes a position information table;

[0041] When two UAVs establish communication, they exchange the latest position information of each UAV obtained by themselves;

[0042] The UAV updates the map according to the latest position information of each UAV obtained through communication.

[0043] In the step S2, the calculation methods of the normalized coverage reward, the normalized fuel consumption penalty, and the normalized repulsion penalty are:

[0044] (1) Calculate the coverage reward J′ 1 , turning penalty J′ 2 , repulsion penalty J′ 3 , and the calculation formula is:

[0045]

[0046] In the formula, (m, n) is the detectable area when the UAV is located at node h; γ h is the weight of node h; the coverable grid when the UAV is located at node h is the set G h ;

[0047]

[0048] represents the steering angle of path s, represents the maximum steering angle;

[0049]

[0050] Cover i (k) is the i-th UAV UAV at time ki Map coverage; D ij For the i-th unmanned aerial vehicle (UAV) i And the j-th unmanned aerial vehicle (UAV) j The distance between; The set of UAVs within the communication range is Q;

[0051] (2) For the coverage return J' 1 And the turning penalty J 2 ', the repulsion penalty J 3 ' are normalized to obtain J 1 , J 2 , J 3 , which are the normalized values of the coverage return, fuel consumption penalty, and repulsion penalty between UAVs.

[0052] In the step S5, the evaluation function of the dynamic window algorithm adopted is:

[0053] G(w) = heading(w)

[0054] In the formula, heading(w) is the azimuth evaluation function, w represents the heading angle, and G(w) represents the evaluation function.

[0055] In the step S5, after guiding the UAV to move forward according to the plan for a fixed duration in combination with the dynamic window algorithm, it further includes the step of updating the map until reaching the end point of the path obtained by the improved A* algorithm.

[0056] In the steps S3 and S5, after updating the map, it further includes the step of calculating the coverage rate. If the coverage rate is greater than the set coverage rate, it directly ends without returning to step S2.

[0057] The present invention has the following beneficial effects compared with the prior art:

[0058] 1. The map model of the present invention can quickly update map information and perform real-time division of different regions. Setting boundary weights effectively reduces the coverage cost caused by regional segmentation. Using the dual-map mode and combining matrix operations can quickly realize map update and judgment of no-fly zones.

[0059] 2. The cooperative search algorithm of the present invention can efficiently guide UAVs to detect uncovered areas, has a high coverage rate, a smooth flight path, and can well avoid obstacles and cope with emergencies.

[0060] 3. The present invention combines image processing technology, the improved A* algorithm, and the dynamic window algorithm, and can quickly guide UAVs to go to uncovered areas to perform tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1Schematic flowchart of a method for multi-UAV collaborative area coverage under communication constraints provided by Embodiment 1 of the present invention;

[0062] Figure 2 Schematic diagram of the predicted path of the UAV;

[0063] Figure 3 UAV no-fly zone avoidance strategy;

[0064] Figure 4 Schematic flowchart of extracting key nodes in the embodiment of the present invention;

[0065] Figure 5 Schematic flowchart of extracting intermediate nodes in the embodiment of the present invention. Detailed implementation manners

[0066] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0067] Embodiment 1

[0068] As Figure 1 shown, Embodiment 1 of the present invention provides a method for multi-UAV collaborative area coverage under communication constraints, including the following steps:

[0069] S1. Map modeling: Set the task area, create two maps and rasterize them at a fixed interval, and assign values to the rasters according to different raster assignment methods to obtain an obstacle avoidance map and a coverage map respectively.

[0070] In the step S1, the raster assignment formula of the obstacle avoidance map Map1 is:

[0071]

[0072] In the formula: Ω 1 is the set of prohibited areas; Ω 2 is the set of other areas; μ mn represents the value of the raster at the m-th row and n-th column of the obstacle avoidance map;

[0073] The raster assignment formula of the coverage map Map2 is:

[0074]

[0075] In the formula: Ω 3 (k) is the set of boundary areas of the undetected area; Ω 4(k) is the set to be searched except for the boundary region; Ω 5 (k) is the set of prohibited regions and searched regions, μ 1mn represents the value of the grid at the m-th row and n-th column of the obstacle avoidance map, k represents the k-th moment, μ 2mn (k) represents the value of the grid at the m-th row and n-th column of the coverage map.

[0076] S2. Use the dynamic window algorithm for path planning, set the evaluation function threshold, and determine whether the maximum value of the evaluation function of each path is greater than the threshold. If it is greater, go to step S3; if it is less, go to step S4.

[0077] Among them, when performing path planning, the evaluation functions adopted by each path include coverage reward, fuel consumption penalty, and repulsive force penalty.

[0078] When performing path planning, it is necessary to establish the UAV motion equation of the unmanned aerial vehicle.

[0079] S3. Select the path with the highest evaluation function value among them, move forward for a fixed duration according to the plan, then update the map according to the position of the unmanned aerial vehicle, and return to step S2;

[0080] In this embodiment, the map is updated according to the position of the unmanned aerial vehicle; as Figure 2 and 3 shown, it is a schematic diagram of the predicted path of the unmanned aerial vehicle and the no-fly zone avoidance strategy.

[0081] S4. Track the unexplored area in the map, binarize the map according to the explored area and other areas to obtain a binary matrix, track the connected areas of the binary matrix, obtain the boundary grid list of the areas of each connected area, select the connected area with an area greater than the area threshold, calculate the distance between the boundary grid and the unmanned aerial vehicle, and select the nearest grid as the target end point;

[0082] S5. Use the improved A* algorithm to calculate the path of the unmanned aerial vehicle to the target end point, and combine the dynamic window algorithm to guide the unmanned aerial vehicle to move forward for a fixed duration according to the plan until it reaches the end point of the path obtained by the improved A* algorithm, and then return to step S2; The use of the improved A* algorithm includes the following steps:

[0083] Use the A* algorithm to find a complete route from the current position of the unmanned aerial vehicle to the target point, and obtain the corresponding node list of this route;

[0084] Extract the key nodes from the node list, and obtain a new node list according to the key nodes;

[0085] Extract and remove the intermediate nodes from the new node list to obtain the final node list;

[0086] Take the final node list as the target point of the dynamic window algorithm in sequence to guide the UAV to fly to the target end point to execute tasks.

[0087] Specifically, in this embodiment, as Figure 4 shown, the specific method for extracting key points in the node list and obtaining a new node list according to the key nodes is as follows:

[0088] Calculate the straight-line equation of the first node and the second node in the node list;

[0089] Substitute the subsequent nodes in sequence and judge whether they conform to the equation. If they conform, continue to judge the next node; if not, judge the previous node as the key node; generate a new straight-line equation based on the key node and the next node of the key node, and repeat the above steps until the end point;

[0090] Record the key nodes to form a new node list.

[0091] Specifically, as Figure 5 shown, the specific method for extracting intermediate nodes in the new node list is as follows:

[0092] Starting from the first node of the new node list, judge in sequence whether there is an obstacle between it and the non-adjacent nodes. If there is no obstacle, the nodes between these two nodes are intermediate nodes, and continue to judge whether there is an obstacle between it and the next node; if there is an obstacle, start from the previous node and continue the above steps until the end point.

[0093] The node list obtained by the traditional A* algorithm has many redundant nodes. In order to optimize the nodes, in this embodiment, the path is first optimized by extracting the key nodes in the list to obtain a new node list. Through the extraction of key points in the previous step, relatively fewer nodes can be obtained, but there are still nodes that can be directly connected between the nodes. By connecting these nodes, the intermediate nodes can be filtered out, thereby maximizing the simplification of the path nodes and shortening the path length. Therefore, the embodiment of the present invention adopts an optimized A* algorithm to optimize the path found by the A* algorithm. Using the new node list as the target point of the dynamic window algorithm in sequence to guide the UAV to fly to the unexplored area to execute tasks.

[0094] Further, in step S5, after guiding the UAV to move forward for a fixed duration according to the plan in combination with the dynamic window algorithm, it further includes the step of updating the map until reaching the end point of the path obtained by the improved A* algorithm.

[0095] In steps S3 and S5, after updating the map, it further includes the step of calculating the coverage rate. If the coverage rate is greater than the set coverage rate, it ends.

[0096] Embodiment Two

[0097] Embodiment 2 of the present invention provides a method for multi-UAV collaborative area coverage under communication constraints. Similar to Embodiment 1, it includes the five steps of S1 to S5. Different from Embodiment 1, the map update method in this embodiment is different. In this embodiment, in order to achieve real-time update of the boundary area, it is assumed that when the UAV is located within a certain grid, it is at the center of the grid; when the grid can be completely detected, it is marked as covered. Specifically, according to the position of the UAV, the following method is used to update the coverage map.

[0098] (1) Obtain the first map map1 and the second map map2 respectively. The grid state of the first map map1 is:

[0099]

[0100] The grid state of the second map map2 is:

[0101]

[0102] Where: Ω 6 (k) is the area outside the detection range; Ω 7 (k) is the area within the detection range; Ω 8 (k) is the boundary set of the area within the detection range; Ω 9 (k) is other areas; μ 3mn and μ 4mn represent the values of the grid at the m-th row and n-th column of the first map and the second map.

[0103] (2) Obtain the matrix map centered on the UAV position in the coverage map, with a length of 2*R e -1; Re represents the radius of the UAV detection range;

[0104] (3) Multiply the corresponding elements of map and map1 to obtain the matrix mapt1;

[0105] (4) Multiply the corresponding elements of the new logicalized matrix of map and map2 to obtain the matrix mapt2;

[0106] (5) Add the matrix mapt1 and the matrix mapt2 to obtain a new map, and assign it to the corresponding position of the coverage map Map2.

[0107] In this embodiment, through the above update method, when the UAV is located within a certain grid, it is regarded as being at the center of the grid; when the grid can be completely detected, it is marked as covered.

[0108] Embodiment 3

[0109] Embodiment 3 of the present invention provides a method for multi-UAV collaborative area coverage under communication constraints. Similar to Embodiment 1, it includes the five steps of S1 to S5. In addition, in step S2 of this embodiment, when performing path planning, a UAV model needs to be established. It is assumed that the unmanned aerial vehicle (UAV) is a particle and flies at a constant speed v at the same altitude in space, and the turning angle range is Its motion equation is:

[0110]

[0111] In the formula: [x i (k), y i (k)] is the position of the i-th UAV at time k; [x i (k + 1), y i (k + 1)] is the position of the i-th UAV at time k + 1; v is the speed of the UAV i ; Δt is the time step; θ i (k) is the heading angle of the i-th UAV at time k.

[0112] Embodiment 4

[0113] Embodiment 4 of the present invention provides a method for multi-UAV collaborative area coverage under communication constraints. Similar to Embodiment 1, it includes the five steps of S1 to S5. Different from Embodiment 1, in steps S2 and S5 of this embodiment, the state transition equation during path planning is specifically:

[0114]

[0115] P i (k) = (x i (k), y i (k), θ i (k)) represents the state information of the i-th UAV i at time k. Within a unit time, the displacement of the UAV is small and it moves in a uniform straight line. The state information P i of the i-th UAV i at time k + 1, P i (k + 1) = (x i (k + 1), y i (k + 1), θ

[0116] In this embodiment, a method of generating a dynamic window is adopted. Using the UAV state transition equation, the possible movement trajectories of the UAV within the time of k - k + N are predicted starting from the k-th moment. The entire path can be regarded as consisting of N straight lines. Among all the paths, the feasible paths are selected, the evaluation function values of all the feasible paths are calculated, the evaluation function is normalized, the path with the highest function value is selected, and this heading angle is taken as the expected value to generate a flight path of a truncated normal distribution. The feasible paths among them are selected, the evaluation function value is calculated again, the evaluation function is normalized, and a path with the highest function value is selected to move forward for a duration of M.

[0117] Embodiment 5

[0118] Embodiment 5 of the present invention provides a method for multi-UAV cooperative area coverage under communication constraints. The same as Embodiment 1, it includes the five steps of S1 to S5. Different from Embodiment 1, in step S2 of this embodiment, the evaluation function adopted by the dynamic window algorithm is:

[0119] J = ω 1 J 1 + ω 2 J 2 + ω 3 J 3 ; (7)

[0120] Wherein, J represents the evaluation function, and J 1 , J 2 , J 3 are the normalized values of the coverage return, fuel consumption penalty, and repulsion penalty between UAVs; ω 1 , ω 2 , ω 3 are the corresponding weights.

[0121] Assume that at the k-th moment, the set of predicted paths of the i-th UAV UAV i is H; the set of nodes of a certain path is H; the set of UAVs within the communication range is Q.

[0122] Among them, the coverage return is that improving the coverage rate is the first task of the UAV, and the calculation formula is:

[0123]

[0124] In the formula, G h represents the detectable area when located at node h; γ h is the weight of node h, and H represents the set of nodes of the path.

[0125] During the search process, the UAV needs more power to turn, consumes a large amount of fuel, and affects the endurance time. Therefore, the calculation formula for the turning penalty is:

[0126]

[0127] Represents the turning angle of path s; Represents the maximum value of the turning angle of the UAV.

[0128] The calculation formula of the repulsive force penalty is:

[0129]

[0130] In the formula, Cover i (k) is the map coverage rate of the UAV i ; D ij is the distance between the i-th UAV i and the j-th UAV j .

[0131] Normalize J 1 ', J 2 ', J 3 ' to obtain J 1 , J 2 , J 3 Then the evaluation function value of this path can be obtained. Under the condition of the maximum J value, according to the path corresponding to the maximum J value, calculate the next heading angle to guide the UAV to move forward for a fixed duration according to the plan.

[0132] Example Five

[0133] Example Five of the present invention provides a method for multi-UAV cooperative area coverage under communication constraints. The same as Example One, it includes the five steps of S1 to S5. Different from Example One, in step S3 of this example, each UAV updates its own map with the position information obtained in communication.

[0134] Specifically, in this example, the communication strategy of each UAV is:

[0135] Each UAV records the positions it has passed and the positions passed by other UAVs learned in communication at each time node, and establishes a position information table;

[0136] When two UAVs establish communication, they exchange the latest position information of each UAV obtained by themselves;

[0137] The UAV updates the map according to the latest position information of each UAV obtained through communication.

[0138] In this example, the UAV i records the positions it has passed and the positions passed by the UAV j (i≠j) learned in communication at each time node, and establishes the position information as shown in Table 1. When the UAV iand UAV j When establishing communication, first the UAV i informs the UAV j of the time node tq of the latest position of the UAV q (q = 1, 2,..., m) it records. If the UAV j has recorded t q and there is subsequent position information of the UAV q , the corresponding time node and position information are returned. The UAV i updates its own map with the position information obtained during communication.

[0139] Table 1 Position Information

[0140]

[0141] Example Six

[0142] Example Five of the present invention provides a method for multi - UAV collaborative area coverage under communication constraints. The same as Example One, it includes the five steps of S1 - S5. Different from Example One, in step S5 of this example, the cost function of the dynamic window algorithm adopted by the A* algorithm is:

[0143] f(n)=g(n)+h(n); (11)

[0144] where f(n) is the total cost of node n; g(n) is the cost from the starting point to node n; h(n) is the cost from node n to the target point, also known as the heuristic function. In this example, h(n) adopts the Euclidean distance, that is:

[0145] h(n)=sqrt((n.x - goal.x)^2+(n.y - goal.y)^2)(12)

[0146] In the formula: n.x and n.y are the horizontal and vertical coordinates of node n respectively; goal.x and goal.y are the horizontal and vertical coordinates of the target point respectively.

[0147] Example Six

[0148] Example Five of the present invention provides a method for multi - UAV collaborative area coverage under communication constraints. The same as Example One, it includes the five steps of S1 - S5. Different from Example One, in step S5, the dynamic window algorithm adopted screens the optimal path through the evaluation function G(w), and the evaluation function is specifically the azimuth evaluation function, that is:

[0149] G(w)=heading(w); (13)

[0150] Wherein, heading(w) is the azimuth evaluation function, w represents the heading angle. By evaluating the angular deviation between the azimuth of the end of the UAV simulation trajectory and the target point, it can promote the UAV to reach the target faster.

[0151] In this embodiment, when the maximum values of the evaluation functions of each path do not satisfy the condition of being greater than the threshold, it indicates that the UAV is in trouble. Through the image processing technology in step S4 and the improved A* algorithm combined with the dynamic window algorithm in step S5, using the condition that the azimuth evaluation function G is the smallest, the heading angle of the UAV is calculated to guide the UAV to perform area coverage.

[0152] The present invention provides a method for multi-UAV collaborative area coverage under communication constraints, which has a high coverage rate, smooth flight paths, can well avoid obstacles and cope with emergencies. By path prediction and setting boundary weights, the coverage cost caused by area segmentation is effectively reduced. Using the dual-map mode and combining matrix operations can quickly realize map update and no-fly zone judgment. By combining image processing technology with the improved A* algorithm and the dynamic window algorithm, the UAV can be quickly guided to the uncovered area to perform tasks.

[0153] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for multi-UAV collaborative area coverage under communication constraints, characterized in that, it includes the following steps: S1. Map modeling: Set the task area, create two maps and rasterize them at a fixed interval, and assign values to the grids according to different grid assignment methods to obtain an obstacle avoidance map and a coverage map respectively; S2. Use the dynamic window algorithm for path planning, set the evaluation function threshold, and judge whether the maximum value of the evaluation function of each path is greater than the threshold. If it is greater, go to step S3; if it is less, go to step S4; When using the dynamic window algorithm for path planning, the evaluation function used for each path is: J = ω 1 J 1 + ω 2 J 2 + ω 3 J 3 ; Among them, J represents the evaluation function, J 1 , J 2 , J 3 are the normalized values of the coverage reward, fuel consumption penalty, and repulsion penalty between UAVs; ω 1 , ω 2 , ω 3 are the corresponding weights; S3. Select the path with the highest evaluation function value, move forward for a fixed duration according to the plan, then update the map according to the position of the UAV, and return to step S2; S4. Track the unexplored areas in the map, binarize the map according to the explored areas and other areas to obtain a binary matrix, track the connected areas of the binary matrix, obtain the boundary grid list of the areas of each connected area, select the connected areas with an area greater than the area threshold, calculate the distance between the boundary grid and the UAV, and select the nearest grid as the target end point; S5. Use the improved A* algorithm to calculate the path from the UAV to the target end point, and combine the dynamic window algorithm to guide the UAV to move forward for a fixed duration according to the plan until it reaches the end point of the path obtained by the improved A* algorithm, and then return to step S2; The improved A* algorithm includes the following steps: Use the A* algorithm to find a complete route from the current position of the UAV to the target point, and obtain the corresponding node list of this route; Extract the key nodes from the node list and obtain a new node list according to the key nodes; Extract and remove the intermediate nodes from the new node list to obtain the final node list; Use the final node list as the target point of the dynamic window algorithm in turn to guide the UAV to fly to the target end point to execute the task.

2. The method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, the specific method for extracting the key points in the node list and obtaining a new node list according to the key nodes is: Calculate the straight line equation of the first node and the second node in the node list; Substitute the subsequent nodes in turn and judge whether they conform to the equation. If they conform, continue to judge the next node; if they do not conform, judge the previous node as the key node; generate a new straight line equation according to the key node and the next node of the key node, and repeat the above steps until the end point; Record the key nodes to form a new node list.

3. The method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, the specific method for extracting the intermediate nodes in the new node list is: Starting from the first node of the new node list, judge in turn whether there is an obstacle between it and the non-adjacent nodes. If there is no obstacle, the nodes between these two nodes are the intermediate nodes, and continue to judge whether there is an obstacle between it and the next node; if there is an obstacle, start from the previous node and continue the above steps until the end point.

4. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, in the step S1, the grid assignment formula of the obstacle avoidance map is: where: Ω 1 is the set of prohibited areas; Ω 2 is the set of other areas; μ 1mn represents the value of the grid at the m-th row and n-th column of the obstacle avoidance map; the grid assignment formula of the coverage map is: Where: Ω 3 (k) is the set of boundary regions of the undetected area; Ω 4 (k) is the set to be searched except for the boundary regions; Ω 5 (k) is the set of prohibited regions and searched regions, μ 2mn (k) represents the value of the grid at the m-th row and n-th column of the coverage map, and k represents the time instant.

5. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, in the step S3, the update method of the coverage map is: (1) Obtain the first map map1 and the second map map2 respectively. The grid state of the first map map1 is: The grid state of the second map map2 is: Where: Ω 6 (k) is the area outside the detection range; Ω 7 (k) is the area within the detection range; Ω 8 (k) is the boundary set of the area within the detection range; Ω 9 (k) is other areas; μ mn represents the value of the grid at the m-th row and n-th column of the obstacle avoidance map; (2) Obtain a matrix map map of length 2*R centered at the UAV position in the coverage map; R e -1; R e represents the radius of the UAV detection range; (3) Multiply the corresponding elements of map and map1 to obtain the matrix mapt1; (4) Multiply the logicalized new matrix of map and the corresponding elements of map2 to obtain the matrix mapt2; (5) Add the matrix mapt1 and the matrix mapt2 to obtain a new map, and assign it to the corresponding position of the coverage map Map2.

6. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, in the step S3, the UAV updates its own map with the position information obtained during communication; the communication strategy of the UAV is: Each UAV records each time node, the positions it has passed and the positions passed by other UAVs learned during communication, and establishes a position information table; When two UAVs establish communication, they exchange the latest position information of each UAV obtained by themselves; The UAV updates the map according to the latest position information of each UAV obtained by communication.

7. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, in the step S2, the calculation methods of the normalized coverage return, the normalized fuel consumption penalty, and the normalized repulsion penalty are: (1) Calculate the coverage return J′ 1 and the turning penalty J′ 2 and the repulsive force penalty J′ 3 , and the calculation formula is as follows: where $(m, n)$ is the detectable area when the UAV is located at node $h$; $\gamma$ h is the weight of node $h$; the coverable grid when the UAV is located at node $h$ is the set $G$ h ; represents the steering angle of path s, represents the maximum steering angle; Cover i (k) is the map coverage rate of the i-th unmanned aerial vehicle (UAV) at time k i ; D ij is the i-th unmanned aerial vehicle (UAV) i and the j-th unmanned aerial vehicle (UAV) j The distance between; the set of UAVs within the communication range is Q; (2) Normalize the coverage reward J′ 1 , the turning penalty J′ 2 , and the repulsion penalty J′ 3 to obtain J 1 , J 2 , J 3 , which are the normalized values of the coverage reward, fuel consumption penalty, and repulsion penalty between UAVs.

8. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, in the step S5, the evaluation function of the dynamic window algorithm adopted is: G(w) = heading(w) wherein, heading(w) is the azimuth evaluation function, w represents the heading angle, and G(w) represents the evaluation function.

9. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, after guiding the UAV to move forward according to the plan for a fixed duration by combining the dynamic window algorithm in the step S5, it further includes the step of updating the map until reaching the end point of the path obtained by the improved A* algorithm.

10. A method for multi-UAV collaborative area coverage under communication constraints according to claim 1, characterized in that, after updating the map in the step S3 and the step S5, it further includes the step of calculating the coverage rate. If the coverage rate is greater than the set coverage rate, it ends.

Images