A method, system, storage medium and terminal for balanced coverage of complex areas of drone clusters based on improved spanning tree

Abstract

The present invention discloses a method, system, storage medium, and terminal for balanced coverage of complex areas by drone swarms based on an improved spanning tree. The method comprises: constructing a structured discrete grid model of the drone swarm coverage area using a grid; adjusting the structured discrete grid model based on a non-uniform adjustable strategy; constructing a collaborative regional coverage model based on the adjusted structured discrete grid model under the premise of satisfying multiple constraints; constructing a spanning tree based on the collaborative regional coverage model, and optimizing the spanning tree to obtain an optimal spanning tree; and constructing a coverage route based on the optimal spanning tree. The present invention can avoid coverage omissions caused by obstacles and irregular regional boundaries in complex areas, and obtains an optimal coverage plan with balanced regional distribution based on an improved spanning tree algorithm under the premise of satisfying multiple constraints, thereby improving coverage completeness and coverage balance while reducing computational complexity.

Description

Technical Field

[0001] The present invention relates to the field of autonomous planning and control of unmanned aerial vehicles (UAVs), and in particular to a method, system, storage medium and terminal for balanced coverage of complex areas of a UAV cluster based on an improved spanning tree. Background Art

[0002] Unmanned swarm coordinated area coverage has widespread applications in both military and civilian fields. In military applications, area coverage search is the primary method for intelligence reconnaissance and information acquisition in future networked battlefields; in civilian applications, area coverage is a means to achieve tasks such as search and rescue and pesticide spraying.

[0003] Due to the complexity of the search area, regional processing technology has become one of the key technologies for collaborative coverage. Although the regional segmentation processing method reduces the complexity of search route planning to a certain extent, it does not take into account the turning process and the existence of regional obstacles. Even with equal area segmentation, it is difficult to achieve balanced coverage within the unmanned swarm. On the other hand, collaborative coverage search planning is an important means of performing regional coverage tasks. It requires planning the optimal coverage route that matches the mission characteristics while meeting the optimal performance indicators. Currently, coverage search planning mostly uses traditional scanning and probability map methods, but there is still a lack of effective unmanned swarm collaborative coverage route planning methods for areas with complex environments.

[0004] Spanning tree-based coverage algorithms are suitable for bounded obstacle environments, but they are prone to "branch blockage" and duplicate coverage. Due to the complex and changing environment and the large number of clusters, conflicts within the unmanned cluster can occur. Furthermore, the regional processing process discards some grids near obstacles to ensure obstacle avoidance, resulting in reduced coverage. Furthermore, spanning tree-based methods require the use of optimization algorithms for solution, but due to multiple constraints such as drone flight constraints, cluster motion coordination constraints, regional boundaries, and internal obstacle constraints, the solution is extremely difficult. Summary of the Invention

[0005] The purpose of the present invention is to address the problems of existing area coverage methods in complex environments such as irregular shapes and internal threats, such as incomplete coverage search, uneven coverage areas, significantly reduced efficiency, and high difficulty in solving problems. The present invention provides a method, system, storage medium and terminal for balanced coverage of complex areas of drone clusters based on an improved spanning tree.

[0006] The object of the present invention is achieved through the following technical solutions:

[0007] In a first solution, a method for balanced coverage of complex areas of a drone cluster based on an improved spanning tree is provided, the method comprising:

[0008] A structured discrete grid model of the drone cluster coverage area is constructed using grids;

[0009] Adjusting structured discrete grid models based on non-uniform adjustable strategies;

[0010] Under the premise of satisfying multiple constraints, a collaborative regional coverage model is constructed based on the adjusted structured discrete grid model;

[0011] Constructing a spanning tree according to the collaborative area coverage model, and optimizing the spanning tree to obtain an optimal spanning tree;

[0012] A covering route is constructed based on the optimal spanning tree.

[0013] In one example, a method for balanced coverage of a complex area by a drone swarm based on an improved spanning tree is provided, wherein a structured discrete grid model of the drone swarm coverage area is constructed using a grid, including:

[0014] Initialize the drone cluster coverage area as a polygon, model the obstacle area within the drone cluster coverage area as a polygon or a circle, wherein the polygon corresponding to the drone cluster coverage area is divided into a uniform grid;

[0015] The grid to be searched is obtained by uniformly discretizing the grids passing through the obstacle area in the direction of the length and width of the minimum circumscribed rectangle of the area.

[0016] The unit grids and their geometric center points corresponding to the obstacle area are removed from the set of grids to be searched and their geometric center points.

[0017] In one example, a method for balanced coverage of complex areas by drone swarms based on an improved spanning tree, wherein a structured discrete grid model is adjusted based on a non-uniform adjustable strategy, includes:

[0018] For the unit grid whose geometric center is inside the obstacle area but only part of it is located in the obstacle area, non-uniform adjustment is performed, wherein a plurality of equally distributed envelope points are generated outside the obstacle area to divide the unit grid again, thereby obtaining small grids that are completely not located in the obstacle envelope area.

[0019] In one example, a method for balanced coverage of complex areas by drone swarms based on an improved spanning tree constructs a coordinated area coverage model based on an adjusted structured discrete grid model, including:

[0020] Assuming that the motion parameters of all drones in the swarm are the same and the lateral detection range is greater than the turning radius, the area coverage model is constructed as follows:

[0021]

[0022]

[0023] Among them, f is the weighted sum of each optimization item, ω1, ω2, ω3 are the weight coefficients of each optimization item, J i represents the range required for UAV i to complete the assigned sub-area coverage mission, J j N represents the range required for UAV j to complete the assigned sub-area coverage mission. u is the number of drones in the unmanned cluster, P i Indicates the number of turns that drone i needs to make, S i The area that each drone needs to cover.

[0024] In one example, a method for balanced coverage of a complex area by a drone swarm based on an improved spanning tree is provided, wherein the spanning tree is constructed according to the collaborative area coverage model, including:

[0025] At the initial moment, N are randomly generated in the unit grid inside the region. g Group root nodes, each group contains N u The root nodes can correspond one-to-one with the initial positions of the drone coverage search, and there is no position conflict between any two root nodes;

[0026] Based on the initial position of the UAV performing the coverage search task, that is, the initial position of the spanning tree root, a depth-first search method is used to construct the spanning tree;

[0027] The initial position of each drone is put into the current solution set, and the u ) The probability of moving from the current node j to the node k (j, k = 1, 2, ..., N s ,j≠k), transition probability The calculation is as follows:

[0028]

[0029] j, k = 1, 2, ..., N s , i=12,…,N u

[0030] in, is the set of nodes connected to node j The total number of nodes in the current spanning tree The difference from the average number of nodes Right now is the number of turns that UAV i covers after transferring to node k, The number of turns that drone i covers on the route before transfer, and must have is the coverage balance evaluation function, which represents the impact of drone i transferring from node j to node k on the overall coverage balance. α and β are adjustment coefficients, α>0; is the route efficiency evaluation function:

[0031]

[0032] In one example, a method for balanced coverage of a complex area by a drone swarm based on an improved spanning tree is provided. The method optimizes the spanning tree to obtain an optimal spanning tree, including:

[0033] The individual transfer probability of each UAV selecting the corresponding candidate node and the average objective function value of the neighbors are calculated, and the node selection results are adjusted to obtain the optimal equilibrium solution for node selection without conflict.

[0034] In one example, a method for balanced coverage of a complex area by a drone swarm based on an improved spanning tree is provided, wherein the method constructs a coverage route based on the optimal spanning tree, including:

[0035] According to the obtained optimal spanning tree, from the reverse extension line of the root node and its subsequent nodes, the distance from the root node is Starting from the position, cover along the optimal spanning tree in a counterclockwise direction, and in the direction of the extension line between the second last node and the last node, the distance from the last node is The optimal spanning tree is covered from the position of until it returns to the starting point, where W represents the lateral detection range of the UAV.

[0036] In the second solution, a complex area balanced coverage system for a drone cluster based on an improved spanning tree is provided, the system comprising:

[0037] Discrete grid model construction module, used to construct a structured discrete grid model of the drone cluster coverage area using grids;

[0038] A discrete grid model non-uniform adjustment module is used to adjust the structured discrete grid model based on a non-uniform adjustable strategy;

[0039] A collaborative regional coverage model construction module is used to construct a collaborative regional coverage model based on the adjusted structured discrete grid model under the premise of satisfying multiple constraints;

[0040] An optimal spanning tree construction module is used to construct a spanning tree according to the collaborative area coverage model and optimize the spanning tree to obtain an optimal spanning tree;

[0041] A route construction module is used to construct a coverage route based on the optimal spanning tree.

[0042] In a third solution, a storage medium is provided, on which computer instructions are stored. When the computer instructions are executed, the steps of the balanced coverage method are executed.

[0043] In a fourth solution, a terminal is provided, including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, and the processor executes the steps of the balanced coverage method when executing the computer instructions.

[0044] It should be further explained that the technical features corresponding to the above options can be combined or replaced with each other to form a new technical solution if there is no conflict.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) The present invention constructs a structured discrete grid model of a complex area with a grid, and adjusts the model based on a non-uniform adjustable strategy to avoid coverage omissions caused by obstacles and irregular regional boundaries in the complex area, and has excellent environmental adaptability. At the same time, under the premise of meeting multiple constraints, a collaborative regional coverage model is constructed based on the adjusted structured discrete grid model, and a tree is generated to find the optimal coverage plan with balanced regional distribution, which improves coverage completeness and coverage balance and reduces computational complexity.

[0047] (2) The present invention uses a grid discretization method along the length and width of the minimum circumscribed rectangle of the region to uniformly discretize the grids passing through the obstacle area, ensuring complete coverage of the search area while minimizing the number of grids to be searched.

[0048] (3) The method proposed in the present invention combines uniform rasterization with non-raster envelope points to perform regional processing, which makes the spanning tree algorithm easy to apply and avoids the coverage omission problem.

[0049] (4) When constructing a spanning tree, the present invention calculates the individual transfer probability of each drone selecting the corresponding candidate node and the average objective function value of the neighbors, adjusts the node selection result, coordinates the node selection conflict based on evolutionary game and obtains the optimal equilibrium solution, comprehensively considers movement safety and coverage efficiency, and achieves better coordination between drones. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A flowchart of a method for balanced coverage of complex areas by drone clusters based on an improved spanning tree according to an embodiment of the present invention is provided;

[0051] Figure 2 A schematic diagram of initial regional settings shown in an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of uniform regional rasterization according to an embodiment of the present invention;

[0053] Figure 4 A schematic diagram of non-uniform regulation according to an embodiment of the present invention;

[0054] Figure 5 A schematic diagram of an initial spanning tree according to an embodiment of the present invention;

[0055] Figure 6 A schematic diagram of an optimal spanning tree according to an embodiment of the present invention;

[0056] Figure 7 This is a schematic diagram of covering route generation according to an embodiment of the present invention.

[0057] Explanation of the numbers in the figure: 1 and 2 indicate no-fly zones, and 3 indicates a radar area. DETAILED DESCRIPTION

[0058] The technical solution of the present invention is described clearly and completely below with reference to the accompanying drawings. It is apparent that the embodiments described are only a portion of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0059] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0060] Figure 1 An embodiment of the present invention provides a method for balanced coverage of complex areas by a drone cluster based on an improved spanning tree, the method comprising:

[0061] A structured discrete grid model of the drone cluster coverage area is constructed using grids;

[0062] Adjusting structured discrete grid models based on non-uniform adjustable strategies;

[0063] Under the premise of satisfying multiple constraints, a collaborative regional coverage model is constructed based on the adjusted structured discrete grid model;

[0064] Constructing a spanning tree according to the collaborative area coverage model, and optimizing the spanning tree to obtain an optimal spanning tree;

[0065] A covering route is constructed based on the optimal spanning tree.

[0066] The present invention is aimed at unmanned swarms performing collaborative regional coverage and reconnaissance tasks. It constructs a structured discrete grid model of complex areas using grids, and adjusts the model based on a non-uniform adjustable strategy to deal with coverage omissions caused by obstacles in the area and irregular regional boundaries. Under the premise of meeting multiple constraints, the optimal coverage plan with balanced regional distribution is obtained based on an improved spanning tree algorithm. The local equilibrium solution is solved to coordinate the conflicts of spanning tree nodes, providing an effective solution for the collaborative regional coverage and reconnaissance planning of unmanned swarms.

[0067] In one example, constructing a structured discrete grid model of the drone cluster coverage area using a grid includes:

[0068] Initialize the drone cluster coverage area as a polygon, model the obstacle area within the drone cluster coverage area as a polygon or a circle, wherein the polygon corresponding to the drone cluster coverage area is divided into a uniform grid;

[0069] The grid to be searched is obtained by uniformly discretizing the grids passing through the obstacle area in the direction of the length and width of the minimum circumscribed rectangle of the area.

[0070] The unit grids and their geometric center points corresponding to the obstacle area are removed from the set of grids to be searched and their geometric center points.

[0071] Specifically, initialize the position of UAV i (x i ,y i ), i=1,2,…,N u , N u is the number of drones in the swarm. All drones fly at a fixed altitude H, and their field of view is simplified to a rectangular area with a horizontal detection range of W and a vertical detection range of L.

[0072] Initialize the drone cluster coverage area S as an n-polygon, whose vertex coordinate set is is the j-th vertex of the polygon, j = 1, 2, ..., n, and the vertices in the vertex coordinate set are arranged in clockwise order.

[0073] Obstacles within the drone swarm coverage area are mainly considered radar and no-fly zones. The radar threat range is modeled as a circle. n=1,2,…,N r ,in is the position of radar n, is the corresponding threat range radius, N r is the number of radars in the area; the no-fly zone is modeled as a polygon, and the no-fly zone m (m=1,2,…,N n ) is the number of edges n m The polygon, then its vertex coordinate set is is the k-th vertex of the polygon, k = 1, 2, ..., n m .

[0074] For some grids that the polygon boundary corresponding to the obstacle area passes through, their geometric centers may not be inside the polygon. In order to ensure a complete coverage search of the area, it is necessary to supplement waypoints outside the area and add the grid to the set of grids to be searched. In order to minimize the number of supplementary grids to be searched (i.e. grids, the same below), the grid discretization method will be adopted along the length and width of the minimum circumscribed rectangle of the area. The polygonal area is rotated at equal intervals within the range of 90° by the equal interval selection search method. By calculating the area of the outer rectangle under different rotations, the minimum circumscribed rectangle can be obtained by comparison. Its length is a, width is b, and the angle between the long side and the horizontal axis is θ. The polygonal area is discretized into a 2W×2W grid along the length and width of the minimum circumscribed rectangle, and the geometric center point of the unit grid that meets the requirements is selected. Nodes for flight route planning, where l = 1, 2, ..., N l , N l is the number of unit grids, symbol Indicates rounding up.

[0075] Furthermore, based on the relevant parameters of the obstacle area (no-fly zone, radar threat zone), based on the regional rasterization results, the unit grids corresponding to the obstacle area and their geometric center points are removed from the set of grids to be searched and their geometric center points. That is, for any unit grid, its relationship with the obstacle area is determined: is it completely within the obstacle area, partially within the obstacle area, or not within the obstacle area at all. If it is completely within the obstacle area, it is removed; if it is not within the obstacle area at all, it is confirmed to be retained; if it is partially within the obstacle area, further adjustments are required.

[0076] In one example, adjusting the structured discrete grid model based on the non-uniform adjustable strategy includes:

[0077] For the unit grid whose geometric center is inside the obstacle area but only part of it is located in the obstacle area, non-uniform adjustment is performed, wherein a plurality of equally distributed envelope points are generated outside the obstacle area to divide the unit grid again, thereby obtaining small grids that are completely not located in the obstacle envelope area.

[0078] Specifically, since the obstacle area also has an unstructured regular shape, the grids that pass through its boundary have a problem where the geometric center is inside the obstacle area, but the unit grid is only partially located in the obstacle area. In order to achieve complete coverage, it is necessary to perform non-uniform adjustment on the unit grid. Generate N b There are equidistantly distributed envelope points, and the interval between them and the obstacle boundary is W / 2.

[0079] The unit grid is further divided into four W×W small grids based on the geometric center. The geometric centers of the small grids that are completely outside the obstacle envelope are obtained and added to the temporary search path point set. After processing all the grids around the obstacle, the two adjacent geometric centers in the set are merged.

[0080] In one example, a collaborative area coverage model is constructed based on the adjusted structured discrete grid model. Before modeling, the constraints are determined. During the cluster collaborative coverage search process, the planning constraints of single-machine coverage and the collaborative constraints of cluster coverage are mainly considered. Among them, the planning constraints of single-machine coverage are: the area S that each drone needs to cover i Must be a connected domain; each sub-region S i It must include the initial position (x i,0 ,y i,0 ), that is, (x i,0 ,y i,0 )∈S i . Collaborative constraint on cluster coverage: The sum of all sub-regions completely covers the entire region S, i.e. S1∪S2∪…∪S Nu =S; the size of the sub-area to be covered by each drone is basically equal, that is, i=1,2,…,N u , where the symbol |·| represents the size of the region; excluding the area around the obstacle, the sub-regions do not overlap, that is, i,j=1,2,…,N u ,i≠j.

[0081] After the coverage area is discretely gridded, the regional coverage problem is converted into a discrete grid coverage problem. Assuming that the motion parameters of all drones in the swarm are the same and the lateral detection distance is greater than the turning radius, the regional coverage model is constructed as follows:

[0082]

[0083]

[0084] Among them, f is the weighted sum of each optimization item, ω1, ω2, ω3 are the weight coefficients of each optimization item, J i represents the range required for UAV i to complete the assigned sub-area coverage mission, J j N represents the range required for UAV j to complete the assigned sub-area coverage mission. u is the number of drones in the unmanned cluster, P i Indicates the number of turns that drone i needs to make, S i The area that each drone needs to cover.

[0085] In one example, constructing a spanning tree according to the collaborative area coverage model includes:

[0086] The construction of the spanning tree depends on its root node. At the initial moment, N cells are randomly generated in the cell grid inside the region. g Group root nodes, each group contains N u The root nodes can correspond one-to-one with the initial positions of the drone coverage search, and there is no position conflict between any two root nodes;

[0087] Based on the initial position of the UAV performing the coverage search task, that is, the initial position of the spanning tree root, a depth-first search method is used to construct the spanning tree;

[0088] The initial position of each drone is put into the current solution set, and the u ) The probability of moving from the current node j to node k (j, k = 1, 2, ..., Ns, j ≠ k), the transfer probability The calculation is as follows:

[0089]

[0090] j, k = 1, 2, ..., N s ,i=1,2,…,N u

[0091] in, is the set of nodes connected to node j The total number of nodes in the current spanning tree The difference from the average number of nodes Right now is the number of turns that UAV i covers after transferring to node k, The number of turns that drone i covers on the route before transfer, and must have is the coverage balance evaluation function, which represents the impact of drone i transferring from node j to node k on the overall coverage balance. α and β are adjustment coefficients, α>0; is the route efficiency evaluation function:

[0092]

[0093] Furthermore, for drones with overlapping nodes to be selected, it is very likely that more than one drone will select the same node. For the sake of both movement safety and coverage efficiency, this situation should be avoided when constructing the spanning tree. Therefore, the optimization of the spanning tree to obtain the optimal spanning tree includes:

[0094] The individual transfer probability of each UAV selecting the corresponding candidate node and the average objective function value of the neighbors are calculated, and the node selection results are adjusted to obtain the optimal equilibrium solution for node selection without conflict.

[0095] Furthermore, constructing a coverage route based on the optimal spanning tree includes:

[0096] According to the obtained optimal spanning tree, from the reverse extension line of the root node and its subsequent nodes, the distance from the root node is Starting from the position, cover along the optimal spanning tree in a counterclockwise direction, and in the direction of the extension line between the second last node and the last node, the distance from the last node is The optimal spanning tree is covered from the position of until it returns to the starting point, where W represents the lateral detection range of the UAV.

[0097] In another embodiment, the present invention provides a complex area balanced coverage system for a drone cluster based on an improved spanning tree, the system comprising:

[0098] Discrete grid model construction module, used to construct a structured discrete grid model of the drone cluster coverage area using grids;

[0099] A discrete grid model non-uniform adjustment module is used to adjust the structured discrete grid model based on a non-uniform adjustable strategy;

[0100] A collaborative regional coverage model construction module is used to construct a collaborative regional coverage model based on the adjusted structured discrete grid model under the premise of satisfying multiple constraints;

[0101] An optimal spanning tree construction module is used to construct a spanning tree according to the collaborative area coverage model and optimize the spanning tree to obtain an optimal spanning tree;

[0102] A route construction module is used to construct a coverage route based on the optimal spanning tree.

[0103] In another embodiment, Figure 2-Figure 7 In this paper, a method for balanced coverage of complex areas by drone swarms based on an improved spanning tree is proposed based on practical situations. The experimental computer configuration is an Intel Core i5-9700F processor with a main frequency of 2.90 GHz and 8 GB of memory, and the software is MATLAB 2019a. The specific steps of this method are as follows:

[0104] Step 1: Initialization

[0105] S11: Initialize the unmanned cluster

[0106] Initialize N u = the positions of 10 drones, N uis the number of drones in the swarm. The drones all fly at a fixed altitude of H = 1000m. Their field of view is simplified to a rectangular area, with a horizontal detection range of W = 420m and a vertical detection range of L = 330m.

[0107] S12: Initialize the coverage area.

[0108] like Figure 2 As shown, the initial coverage area is a rectangle, and its vertex coordinate set is

[0109] The vertices in the vertex coordinate set are arranged in clockwise order. Obstacles in the area are mainly considered radar and no-fly zones. The threat range of the radar is modeled as a circle The position of radar n is The corresponding threat range radius is The number of radars in the area is N r =1; the no-fly zone is modeled as a polygon, and the number of no-fly zones within the area is N n =2, no-fly zones 1 and 2 are both polygons with 5 sides, and the vertex coordinate set of no-fly zone 1 is

[0110] The vertex coordinate set of no-fly zone 2 is

[0111] Step 2: Discrete rasterization of complex areas

[0112] S21: Uniform grid discretization of the region

[0113] For some grids that the polygonal area boundary passes through, their geometric centers may not be inside the polygon. To ensure complete coverage of the area, it is necessary to add waypoints outside the area and add the grid to the set of grids to be searched.

[0114] To minimize the number of grid cells to be searched, a grid discretization method was used along the length and width of the minimum bounding rectangle of the region. Using an equal-interval selection search method, the polygonal region was rotated at equal intervals within a range of 90°. By calculating the areas of the bounding rectangles under different rotations, a minimum bounding rectangle was obtained: a length of 15 km, a width of 10 km, and an angle of θ = 60° between the long side and the horizontal axis.

[0115] like Figure 3 As shown, the polygonal area is discretized into 840m×840m grids along the length and width of the minimum circumscribed rectangle, and the geometric center point of the unit grid that meets the requirements is selected. Nodes for flight route planning, where l = 1, 2, ..., N l , N l is the number of unit grids, symbol Indicates rounding up.

[0116] S22: Eliminate cell grids in the obstacle area

[0117] According to the relevant parameters of the obstacle area (no-fly zone, radar threat zone), based on the regional rasterization results in S21, the unit grids corresponding to the obstacle area and their geometric center points are removed from the set of grids to be searched and their geometric center points. That is, for any unit grid, its relationship with the obstacle area is determined: completely within the obstacle area, partially within the obstacle area, or not within the obstacle area at all. If it is completely within the obstacle area, it is removed; if it is not within the obstacle area at all, it is confirmed to be retained; if it is partially within the obstacle area, further adjustment is required.

[0118] S23: Non-uniform adjustment of the grid around the obstacle

[0119] Since the obstacle area also has an unstructured regular shape, the boundary of the grid passes through some of the grids, and the geometric center is inside the obstacle area, but the unit grid is only partially located in the obstacle area. In order to achieve complete coverage, it is necessary to perform non-uniform adjustment on the unit grid. Generate N b = 5 equally spaced envelope points, with a distance from the obstacle boundary of W / 2 = 210m.

[0120] like Figure 4 As shown, the unit grid is further divided into four small grids of W×W=210m×210m based on the geometric center. The geometric centers of the small grids that are completely outside the obstacle envelope are obtained and added to the temporary search path point set. After processing all the grids surrounding the obstacle, the two adjacent geometric centers in the set are merged.

[0121] Step 3: Build an unmanned cluster collaborative area coverage model

[0122] S31: Determination of constraints

[0123] In the cluster collaborative coverage search process, the planning constraints of single-machine coverage and the collaborative constraints of cluster coverage are mainly considered.

[0124] Planning constraints for single-machine coverage: The area S that each drone needs to cover i Must be a connected domain; each sub-region S i Must include the initial position of the corresponding drone performing the coverage mission Right now

[0125] The synergy constraint of cluster coverage is that the sum of all sub-areas completely covers the entire area S, i.e. The size of the sub-area to be covered by each drone is basically the same, that is, i=1,2,…,N u , where the symbol |·| represents the size of the region; excluding the area around the obstacle, the sub-regions do not overlap, that is, i,j=1,2,…,N u ,i≠j.

[0126] S32: Regional coverage model construction

[0127] After the coverage area is discretely gridded, the regional coverage problem is converted into a discrete grid coverage problem. Assuming that the motion parameters of all drones in the swarm are the same and the lateral detection distance is greater than the turning radius, the regional coverage model can be constructed as follows:

[0128]

[0129]

[0130] Among them, J i P represents the distance required for UAV i to complete the assigned sub-area coverage task, which is the sum of the length of the coverage route and the distance from the current position to the initial position of the coverage task; i represents the number of turns that UAV i needs to make, ω1, ω2, and ω3 are the weight coefficients of each optimization item, ω1 = 0.01, ω2 = 0.01, ω3 = 5, and f is the weighted sum of each optimization item.

[0131] Step 4: Construction and optimization of the spanning tree

[0132] S41: Generate the root node of the spanning tree

[0133] The construction of the spanning tree depends on its root node. At the initial moment, N cells are randomly generated in the cell grid inside the region. g = 30 groups of root nodes, each group contains N u = 10 root nodes, which can correspond to the initial positions of the drone coverage search one by one, and no two root nodes will conflict. The initial spanning tree is as follows Figure 5 shown.

[0134] S42: Parallel construction of spanning trees

[0135] Based on the initial position of the UAV performing the coverage search task, that is, the initial position of the spanning tree root, the spanning tree is constructed using the depth-first search method. The initial position of each UAV is placed in the current solution set, and the initial position of each UAV is calculated byu ) The probability of moving from the current node j to the node k (j, k = 1, 2, ..., N s ,j≠k), transition probability The calculation is as follows:

[0136]

[0137] j, k = 1, 2, ..., N s , i=1,2,…,N u

[0138] in, is the set of nodes connected to spanning tree node j The total number of nodes in the current spanning tree The difference from the average number of nodes Right now is the number of turns that UAV i covers after transferring to node k, The number of turns that drone i covers on the route before transfer, and must have is the coverage balance evaluation function, which represents the impact of drone i transferring from node j to node k on the overall coverage balance. α and β are adjustment coefficients, α = 10, β = 0; is the route efficiency evaluation function, which aims to guide UAV i to preferentially move to the node with fewer turns.

[0139] S43: Spanning Tree Conflict Coordination

[0140] For drones with overlapping candidate nodes, it is very likely that more than one drone will select the same node. For the sake of both motion safety and coverage efficiency, this situation should be avoided when constructing the spanning tree. The individual transition probability of each drone selecting the corresponding candidate node and the average objective function value of its neighbors are calculated, and the node selection results are adjusted to obtain the optimal equilibrium solution for conflict-free node selection.

[0141] Step 5: Construct coverage routes based on the optimal spanning tree

[0142] The optimal spanning tree is Figure 6 As shown, according to the obtained optimal spanning tree, from the root node and its subsequent nodes on the reverse extension line, the distance from the root node is Starting from the position, cover along the optimal spanning tree in a counterclockwise direction, and in the direction of the extension line between the second last node and the last node, the distance from the last node is Start from the position of the optimal spanning tree and cover it in the reverse direction until you return to the starting point. Based on the optimal spanning tree, generate Figure 7The coverage routes shown in the figure are represented by black solid lines, where each closed line corresponds to the coverage search route of a UAV. The coverage search routes of different UAVs do not conflict with each other. Figure 5 , the optimized spanning tree node branches (such as Figure 5 The upper right corner) is reduced, and the corresponding coverage search route does not require too many turns, and the cost is lower.

[0143] In another example, the present invention provides a storage medium having computer instructions stored thereon, wherein the computer instructions execute the balanced coverage method when executed.

[0144] Based on this understanding, the technical solution of this embodiment, or the portion that contributes to the prior art, or the portion 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 a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing 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.

[0145] In another example, the present invention provides a terminal including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, and the processor executes the balanced coverage method when executing the computer instructions.

[0146] The processor may be a single-core or multi-core central processing unit or a specific integrated circuit, or one or more integrated circuits configured to implement the present invention.

[0147] Embodiments of the subject matter and functional operations described in this specification may be implemented in: tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or a combination of one or more thereof. Embodiments of the subject matter described in this specification may be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier for execution by a data processing apparatus or to control the operation of the data processing apparatus. Alternatively or in addition, the program instructions may be encoded on an artificially generated propagated signal, such as a machine-generated electrical, optical, or electromagnetic signal, that is generated to encode and transmit information to a suitable receiver apparatus for execution by the data processing apparatus.

[0148] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform the corresponding functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can be implemented as, special purpose logic circuitry, such as an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

[0149] Processors suitable for executing computer programs include, for example, general-purpose and / or special-purpose microprocessors, or any other type of central processing unit. Typically, a central processing unit will receive instructions and data from a read-only memory and / or random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or the computer will be operably coupled to such a mass storage device to receive data from it or to transmit data to it, or both. However, a computer does not necessarily have such a device. In addition, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name a few.

[0150] Although this specification includes many specific implementation details, these should not be interpreted as limiting the scope of any invention or the scope of protection claimed, but are mainly used to describe the features of specific embodiments of specific inventions. Certain features described in multiple embodiments within this specification may also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment may also be implemented separately in multiple embodiments or in any suitable sub-combination. In addition, although features may work in certain combinations as described above and even initially claimed as such, one or more features from the claimed combination may be removed from the combination in some cases, and the claimed combination may point to a sub-combination or a variation of the sub-combination.

[0151] Similarly, although operations are depicted in a particular order in the accompanying drawings, this should not be understood as requiring that these operations be performed in the particular order shown or performed sequentially, or that all illustrated operations be performed to achieve the desired results. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of various system modules and components in the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product, or packaged into multiple software products.

[0152] The above specific implementation methods are detailed descriptions of the present invention. It cannot be considered that the specific implementation methods of the present invention are limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, they can make several simple deductions and substitutions without departing from the concept of the present invention, which should be regarded as falling within the scope of protection of the present invention.

Claims

1. A method for balanced coverage of complex areas by drone clusters based on an improved spanning tree, characterized in that: The method comprises: A structured discrete grid model of the drone cluster coverage area is constructed using grids; Adjusting structured discrete grid models based on non-uniform adjustable strategies; Under the premise of satisfying multiple constraints, a collaborative regional coverage model is constructed according to the adjusted structured discrete grid model; the collaborative regional coverage model is constructed according to the adjusted structured discrete grid model, including: Assuming that the motion parameters of all drones in the swarm are the same and the lateral detection range is greater than the turning radius, the area coverage model is constructed as follows: ,in, is the weighted sum of each optimization item, 、 、 are the weight coefficients of each optimization item, It represents the range required for UAV i to complete the assigned sub-area coverage mission, represents the range required for UAV j to complete the assigned sub-area coverage mission, is the number of drones in the unmanned cluster, Indicates the number of times drone i needs to turn, The area that each drone needs to cover; Constructing a spanning tree according to the collaborative area coverage model and optimizing the spanning tree to obtain an optimal spanning tree; constructing the spanning tree according to the collaborative area coverage model includes: At the initial moment, the cells are randomly generated in the cell grid inside the region. Group root node, each group contains root nodes, which can correspond one-to-one with the initial positions of the drone coverage search, and there is no position conflict between any two root nodes; Based on the initial position of the UAV performing the coverage search task, that is, the initial position of the spanning tree root, a depth-first search method is used to construct the spanning tree; The initial position of each drone is placed in the current solution set, and the The probability of moving from the current node j to node k , transition probability The calculation is as follows: ,in, is the set of nodes connected to node j , The total number of nodes in the current spanning tree The difference from the average number of nodes ,Right now ; is the number of turns that UAV i covers after transferring to node k, The number of turns that UAV II covers on the route before transfer, and must have ; is the coverage balance evaluation function, which represents the impact of drone i transferring from node j to node k on the overall coverage balance. , and is the adjustment coefficient, ; is the route efficiency evaluation function: ； A covering route is constructed based on the optimal spanning tree.

2. The method for balanced coverage of complex areas by drone clusters based on an improved spanning tree according to claim 1 is characterized in that: The structured discrete grid model of the drone cluster coverage area is constructed using a grid, including: Initialize the drone cluster coverage area as a polygon, model the obstacle area within the drone cluster coverage area as a polygon or a circle, wherein the polygon corresponding to the drone cluster coverage area is divided into a uniform grid; The grid to be searched is obtained by uniformly discretizing the grids passing through the obstacle area in the direction of the length and width of the minimum circumscribed rectangle of the area. The unit grids and their geometric center points corresponding to the obstacle area are removed from the set of grids to be searched and their geometric center points.

3. The method for balanced coverage of complex areas by drone clusters based on an improved spanning tree according to claim 2 is characterized in that: The adjusting of the structured discrete grid model based on the non-uniform adjustable strategy includes: For the unit grid whose geometric center is inside the obstacle area but only part of it is located in the obstacle area, non-uniform adjustment is performed, wherein a plurality of equally distributed envelope points are generated outside the obstacle area to divide the unit grid again, thereby obtaining small grids that are completely not located in the obstacle envelope area.

4. The method for balanced coverage of complex areas by drone clusters based on an improved spanning tree according to claim 1 is characterized in that: Optimizing the spanning tree to obtain the optimal spanning tree includes: The individual transfer probability of each UAV selecting the corresponding candidate node and the average objective function value of the neighbors are calculated, and the node selection results are adjusted to obtain the optimal equilibrium solution for node selection without conflict.

5. The method for balanced coverage of complex areas by drone clusters based on an improved spanning tree according to claim 1 is characterized in that: The constructing of a coverage route based on the optimal spanning tree includes: According to the obtained optimal spanning tree, from the reverse extension line of the root node and its subsequent nodes, the distance from the root node is Starting from the position, cover along the optimal spanning tree in a counterclockwise direction, and in the direction of the extension line between the second last node and the last node, the distance from the last node is The optimal spanning tree is covered from the position of until it returns to the starting point, where W represents the lateral detection range of the UAV.

6. A complex area balanced coverage system for drone swarms based on an improved spanning tree, characterized by: The system comprises: Discrete grid model construction module, used to construct a structured discrete grid model of the drone cluster coverage area using grids; A discrete grid model non-uniform adjustment module is used to adjust the structured discrete grid model based on a non-uniform adjustable strategy; The collaborative regional coverage model construction module is used to construct the collaborative regional coverage model based on the adjusted structured discrete grid model under the premise of satisfying multiple constraints; the construction of the collaborative regional coverage model based on the adjusted structured discrete grid model includes: Assuming that the motion parameters of all drones in the swarm are the same and the lateral detection range is greater than the turning radius, the area coverage model is constructed as follows: ,in, is the weighted sum of each optimization item, 、 、 are the weight coefficients of each optimization item, It represents the range required for UAV i to complete the assigned sub-area coverage mission, represents the range required for UAV j to complete the assigned sub-area coverage mission, is the number of drones in the unmanned cluster, Indicates the number of times drone i needs to turn, The area that each drone needs to cover; An optimal spanning tree construction module is configured to construct a spanning tree according to the collaborative area coverage model and optimize the spanning tree to obtain an optimal spanning tree; the construction of the spanning tree according to the collaborative area coverage model includes: At the initial moment, the cells are randomly generated in the cell grid inside the region. Group root node, each group contains root nodes, which can correspond one-to-one with the initial positions of the drone coverage search, and there is no position conflict between any two root nodes; Based on the initial position of the UAV performing the coverage search task, that is, the initial position of the spanning tree root, a depth-first search method is used to construct the spanning tree; The initial position of each drone is placed in the current solution set, and the The probability of moving from the current node j to node k , transition probability The calculation is as follows: ,in, is the set of nodes connected to node j , The total number of nodes in the current spanning tree The difference from the average number of nodes ,Right now ; is the number of turns that UAV i covers after transferring to node k, The number of turns that UAV II covers on the route before transfer, and must have ; is the coverage balance evaluation function, which represents the impact of drone i transferring from node j to node k on the overall coverage balance. , and is the adjustment coefficient, ; is the route efficiency evaluation function: ； A route construction module is used to construct a coverage route based on the optimal spanning tree.

7. A storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed, the balanced coverage method according to any one of claims 1 to 5 is executed.

8. A terminal comprising a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, characterized in that: When the processor runs the computer instructions, the balanced coverage method according to any one of claims 1 to 5 is executed.

Images