Large-scale unmanned vehicle type conversion method and device based on descending order strategy

Abstract

The application discloses a large-scale unmanned aerial vehicle configuration transformation method and device based on a descending order strategy, which comprises the following steps: arranging the unmanned aerial vehicle cluster in an original configuration in descending order based on the node distance between each starting node and a reference point in the original configuration; arranging the unmanned aerial vehicle cluster in a target configuration in descending order in the same way; sequentially performing global pairing on all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence to generate a plurality of pairs; arranging the plurality of pairs in descending order according to an ideal path, and performing pairing optimization processing on the preselected pair sequence obtained by the descending order arrangement to obtain a quasi-pair sequence; and performing path obstacle avoidance processing on each pair in the quasi-pair sequence to generate a path plan of the unmanned aerial vehicle configuration. Thus, the computational complexity can be reduced from O(N!) to O(N 2 +δM 2 +L×N), M, L << N, and the energy consumption of the unmanned aerial vehicle is balanced on the basis of ensuring the optimal path, thereby improving the management efficiency of the large-scale unmanned aerial vehicle cluster.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) technology, and particularly relates to a method and apparatus for large-scale unmanned vehicle configuration transformation based on a descending order strategy. Background Technology

[0002] The rapid development of UAV technology has led to the widespread application of UAVs in military, film and television aerial photography, environmental monitoring, geographic surveying, power line inspection, and other civilian fields. Faced with increasingly complex mission requirements, single UAVs are insufficient to complete tasks independently, thus giving rise to multi-UAV collaborative mission execution. UAV path planning is the core of UAV collaborative control, making research on multi-UAV collaborative path planning of great significance. Multi-UAV collaborative path planning essentially coordinates the paths between multiple UAVs to complete collaborative tasks. Existing path planning methods based on graph search and sampling exhibit high efficiency and excellent global optimization performance in environmental search; however, these algorithms suffer from dynamic obstacle avoidance and communication problems in large-scale environments. Heuristic path planning methods, such as ant colony optimization, particle swarm optimization, bee colony optimization, genetic algorithms, and simulated annealing, can effectively plan paths and allocate tasks in complex, dynamic, and multi-constraint environments, but they have limitations such as high computational cost and susceptibility to local optima. Learning-based algorithms can adjust their strategies by continuously learning new information from the environment, making them highly adaptable and suitable for dynamic and uncertain environments. However, learning algorithms are heavily dependent on large-scale, high-quality training data. When the data is insufficient or inaccurate, the performance of the algorithm may drop significantly. Furthermore, the algorithm has poor interpretability, which is not conducive to system debugging and optimization.

[0003] Existing multi-UAV cooperative path planning algorithms primarily address dynamic obstacle avoidance and search coverage problems, with limited coverage of multi-UAV swarm configuration transformations. The computational complexity of existing algorithms increases exponentially with swarm size, making global task planning unsuitable for large-scale UAV swarm configuration transformations. Furthermore, general optimization methods often only consider factors such as shortest path allocation and energy minimization, neglecting path standard deviation. This leads to uneven energy consumption among UAVs, increasing the difficulty of managing UAV swarms. Summary of the Invention

[0004] This invention provides a method and apparatus for large-scale unmanned aerial vehicle (UAV) configuration transformation based on a descending order strategy. The method can rationally plan the paths between each pair, thereby achieving the goal of minimizing the paths of each pair and balancing energy consumption, thus improving the management efficiency of large-scale UAV swarms.

[0005] According to a first aspect of the present invention, a method for large-scale unmanned aerial vehicle (UAV) configuration transformation based on a descending order strategy is provided. The method includes: arranging a UAV cluster in the original configuration in descending order based on the node distances from each starting node to a reference point in the original configuration, generating an original configuration sequence; arranging a UAV cluster in the target configuration in descending order based on the node distances from each target node to a reference point in the target configuration corresponding to the original configuration, generating a target configuration sequence; wherein the starting node is used to indicate the azimuth information of the UAV in the original configuration; the target node is used to indicate the azimuth information of the UAV in the target configuration; performing global pairing sequentially on all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence to generate several pairs; arranging the several pairs in descending order according to the ideal paths corresponding to the pairs, and performing pairing optimization processing on the pre-selected pairing sequence obtained by the descending order arrangement to obtain a quasi-pairing sequence; and performing path obstacle avoidance processing on each pair in the quasi-pairing sequence to generate a path plan for the UAV configuration.

[0006] According to a second aspect of the present invention, a large-scale unmanned aerial vehicle (UAV) configuration transformation device based on a descending order strategy is also provided. The device includes: a first generation module, configured to arrange the UAV cluster in the original configuration in descending order based on the node distance between each starting node and a reference point in the original configuration, generating an original configuration sequence; a second generation module, configured to arrange the UAV cluster in the target configuration in descending order based on the node distance between each target node and a reference point in the target configuration corresponding to the original configuration, generating a target configuration sequence; wherein the starting node is used to indicate the azimuth information of the UAV in the original configuration; the target node is used to indicate the azimuth information of the UAV in the target configuration; a global pairing module, configured to sequentially perform global pairing on all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence, generating several pairings; an optimization processing module, configured to sort the several pairings in descending order according to the ideal paths corresponding to the pairings, and perform pairing optimization processing on the obtained pre-selected pairing sequence to obtain a quasi-pairing sequence; and a path obstacle avoidance processing module, configured to perform path obstacle avoidance processing on each pairing in the quasi-pairing sequence to generate a path plan for the UAV configuration.

[0007] According to a third aspect of the present invention, an electronic device is also provided, comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method as described in the first aspect.

[0008] According to a fourth aspect of the present invention, a computer-readable medium is also provided, on which a computer program is stored, wherein the program, when executed by a processor, implements the method described in the first aspect.

[0009] This invention provides a method for large-scale unmanned aerial vehicle (UAV) configuration transformation based on a descending order strategy. The method includes: first, arranging the UAV cluster in the original configuration in descending order based on the node distances from each starting node to a reference point in the original configuration, generating an original configuration sequence; then, arranging the UAV cluster in the target configuration in descending order based on the node distances from each target node to a reference point in the target configuration corresponding to the original configuration, generating a target configuration sequence; second, globally pairing all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence sequentially, generating several pairs; then, arranging the pairs in descending order according to the ideal paths corresponding to the pairs, and performing pairing optimization processing on the pre-selected pairing sequences obtained by the descending order arrangement to obtain a quasi-pairing sequence; finally, performing path obstacle avoidance processing on each pair in the quasi-pairing sequence to generate a path plan for the UAV configuration. This embodiment employs a descending order strategy for the transformation of large-scale UAV configurations, combining global pairing, local optimization, individual node optimization, and continuous steps of path obstacle avoidance processing; thus, it can reduce the computational complexity from O(N!) to O(N...). 2 +δM 2 +L×N), M,L<<N; Moreover, it can achieve balanced energy consumption of UAVs while ensuring optimal path, thereby improving the management efficiency of large-scale UAV swarms.

[0010] The following sections will describe some specific embodiments of the invention in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings:

[0011] Figure 1 This is a flowchart illustrating a large-scale unmanned structure transformation method based on a descending order strategy, as provided in an embodiment of the present invention.

[0012] Figure 2 This is a schematic diagram of the process for generating quasi-paired sequences provided in one embodiment of the present invention;

[0013] Figure 3 This is a schematic diagram of a path planning process for generating unmanned vehicle configurations provided in one embodiment of the present invention;

[0014] Figure 4 This is a schematic diagram of a large-scale unmanned structure transformation device based on a descending order strategy, provided as an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0016] A primitive drone swarm configuration typically refers to the state of a drone swarm as it naturally forms or follows a basic assembly pattern before initial deployment or before executing a specific mission plan. It is generally based on simple networking and distribution, without specific formation or structural design for a particular mission. A target-configuration drone swarm configuration, on the other hand, is a carefully planned and designed formation, structure, or layout that the drone swarm aims to achieve to maximize mission performance, based on specific mission requirements, environmental conditions, and tactical needs.

[0017] This embodiment addresses the technical problem of path planning for transforming a drone swarm from its original configuration to a target configuration. For example... Figure 1 The diagram shown is a flowchart illustrating a large-scale unmanned structure transformation method based on a descending order strategy, according to an embodiment of the present invention.

[0018] A method for large-scale unmanned vehicle configuration transformation based on a descending order strategy includes at least the following steps:

[0019] S101, Based on the node distance between each starting node and the reference point in the original configuration, the UAV cluster in the original configuration is sorted in descending order to generate the original configuration sequence;

[0020] S102, based on the node distance between each target node and the reference point in the target configuration corresponding to the original configuration, the UAV cluster in the target configuration is sorted in descending order to generate a target configuration sequence; wherein, the starting node is used to indicate the orientation information of the UAV in the original configuration; the target node is used to indicate the orientation information of the UAV in the target configuration;

[0021] S103, perform global pairing of all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence in sequence to generate several pairs;

[0022] S104, sort the several pairs in descending order according to the ideal path corresponding to the pair, and perform pairing optimization processing on the pre-selected pairing sequence obtained by descending order to obtain the quasi-pairing sequence;

[0023] S105, perform path obstacle avoidance processing on each pair in the alignment sequence to generate path planning for the unmanned machine type.

[0024] In S101 and S102, the reference point can be the centroid of the original configuration, the centroid of the target configuration, or other reference points. Using the reference point as a benchmark, nodes are sorted by distance; this transforms the spatial distribution of the UAV swarm into quantifiable sequential data (such as distance vectors), providing a unified evaluation benchmark for both the original and target configurations.

[0025] Here, location information includes direction information and position information.

[0026] For example, based on the position information of each UAV in the target configuration, the centroid of the target configuration is determined; the node distance from each starting node in the original configuration to the centroid is determined; based on the node distances corresponding to the starting nodes, the UAV cluster in the original configuration is sorted in descending order to generate an original configuration sequence; the node distance from each target node in the target configuration to the centroid is determined; based on the node distances corresponding to the target nodes, the UAV cluster in the target configuration is sorted in descending order to generate a target configuration sequence.

[0027] For example: Step 1.1: Calculate the centroid c of the target configuration g UAV swarm. g ;

[0028]

[0029] Step 1.2: Using the centroid c of the target configuration UAV swarm g Using the origin as the reference point, calculate the distance from each UAV node in the original configuration s to the centroid c of the target configuration. g The distance between nodes l s,j All drone nodes in the drone cluster are reordered according to the principle of descending distance.

[0030]

[0031] Step 1.3: Using the centroid c of the target configuration UAV swarm g Using the origin as the reference point, calculate the distance from each UAV node in the target configuration g to the centroid c of the target configuration. g The distance between nodes l g,i All drone nodes in the drone cluster are reordered according to the principle of descending distance.

[0032]

[0033] In S103, based on preset rules or algorithms, all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence are globally paired sequentially to generate several pairs;

[0034] For example, the drone clusters in the original configuration sequence and the drone clusters in the target configuration sequence are paired according to the sequence order, and the optimal pairing is selected based on a greedy strategy, generating several pairings. This enables fast and uniform path planning for drone nodes in both the original and target configurations.

[0035] Specifically, all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence are globally paired sequentially to generate several pairs; including:

[0036] S1, select the first starting node from the original configuration sequence and the first target node from the target configuration sequence;

[0037] S2, If the distance between the first starting node and the first target node is greater than the distance between the first target node, then determine the path length between the first starting node and each target node in the target configuration sequence, and select the target node with the shortest path length from the target configuration as the pairing node of the first starting node.

[0038] S3. If the distance between the first starting node and the first target node is less than the distance between the first target node and the first target node, then determine the path length between the first target node and each starting node in the original configuration sequence, and select the starting node with the shortest path length from the original configuration as the pairing node of the first target node.

[0039] S4, Remove successfully paired nodes from the original configuration sequence and the target configuration sequence respectively, and generate the first update sequence corresponding to the original configuration sequence and the second update sequence corresponding to the target configuration sequence;

[0040] S5. Repeat steps S1 to S4 until all UAVs in the original configuration sequence or the target configuration sequence are paired, generating several pairs.

[0041] For example: Step 2.1: The node distance between the k-th starting node in the original configuration sequence and the centroid of the target configuration is l. s,k The node distance between the k-th target node in the target configuration sequence and the centroid of the target configuration is l. g,k If l s,k ≥l g,k If so, proceed to step 2.2; if l s,k ≤l g,k Then proceed to step 2.3;

[0042] Step 2.2: Calculate the starting node s k The path length between each target node is determined by a greedy strategy, which selects the target node with the shortest path length and removes the successfully paired start and target nodes from the original configuration s and the target configuration g.

[0043] Step 2.3: Calculate the target node g k The path length between each starting node is determined by a greedy strategy, which selects the starting node with the shortest path length. The successfully paired starting and target nodes are then removed from the original configuration s and the target configuration g.

[0044] Step 2.4: Repeat steps 2.1-2.3 until all drones are paired;

[0045] In S104, based on preset rules or models, the pre-selected pairing sequences obtained by descending order are subjected to pairing optimization processing to obtain quasi-pairing sequences.

[0046] For example, several pairs are arranged in descending order according to the ideal paths corresponding to the pairs, and the pre-selected pairing sequence obtained by the descending order is subjected to pairing optimization processing to obtain a quasi-pairing sequence; including: determining the ideal path corresponding to each pair in the several pairs, and sorting the several pairs according to the descending order of the ideal paths to obtain a pre-selected pairing sequence; selecting the first N pairs from the pre-selected pairing sequence by random sampling of nodes, and performing local optimization processing on the subsequences formed by the first N pairs to generate an optimized pairing sequence; and performing individual optimization processing on the optimized pairing sequence to obtain a quasi-pairing sequence.

[0047] This embodiment first performs local optimization on the pre-selected pairing sequence, and then performs individual optimization. This not only reduces the average path length of the UAVs, but also reduces the standard deviation of the UAV path length, thereby balancing the energy consumption of UAVs in different configurations and improving the management efficiency of UAV swarms.

[0048] In S105, based on the improved A* algorithm, dynamic programming algorithm, or artificial potential field method, path obstacle avoidance processing is performed on each pair in the alignment pairing sequence to generate path planning for unmanned structures.

[0049] This embodiment employs a descending order strategy for transforming large-scale unmanned structures, combining global pairing, local optimization, individual optimization, and sequential steps of path obstacle avoidance. This not only reduces the computational complexity from O(N!) to O(N...) 2 +δM 2 +L×N), M,L<<N, and achieves path minimization and drone energy balance; thereby improving the management efficiency of large-scale drone swarms.

[0050] In a preferred embodiment of this example, the original configuration and the target configuration are obtained by the following method: extracting the outer surface contour points from the three-dimensional model corresponding to the first unmanned vehicle configuration, and scaling the extracted outer surface contour points to generate the original configuration; obtaining the second unmanned vehicle configuration corresponding to the first unmanned vehicle configuration; extracting the outer surface contour points from the three-dimensional model corresponding to the second unmanned vehicle configuration, and scaling the extracted outer surface contour points to generate the target configuration.

[0051] Therefore, by appropriately scaling different configurations of UAV swarms, the safety of UAV flight and the feasibility of path planning can be ensured while reducing energy consumption through the synergistic effects of aerodynamic drag reduction, power system load reduction, and control response acceleration.

[0052] like Figure 2 The diagram shown is a schematic flowchart of the process for generating quasi-paired sequences provided in one embodiment of the present invention.

[0053] Generating quasi-paired sequences includes at least the following steps:

[0054] S201, determine the ideal path for each pair in a number of pairings, and sort the pairings in descending order of the ideal paths to obtain a pre-selected pairing sequence;

[0055] S202, Select the first N pairs from the pre-selected pairing sequence by random sampling of nodes, and perform local optimization processing on the subsequences formed by the first N pairs to generate the optimized pairing sequence;

[0056] S203, perform single-unit optimization on the optimized paired sequences to obtain quasi-paired sequences.

[0057] In S201, for any pair among several pairings: obtain the starting node corresponding to the original configuration and the target node corresponding to the target configuration in the pairing; determine the ideal path corresponding to the pairing based on the coordinate information corresponding to the starting node and the coordinate information corresponding to the target node; and arrange the several pairings in descending order based on the ideal path to generate a pre-selected pairing sequence.

[0058] In S202, the first N pairs are selected from the pre-selected pairing sequence by random sampling of nodes, and the subsequences formed by the first N pairs are locally optimized based on preset rules or models to generate the optimized pairing sequence.

[0059] Specifically, the first N pairs are selected from the pre-selected pairing sequence using a node random sampling method, and the subsequences formed by the first N pairs are locally optimized to generate an optimized pairing sequence; including:

[0060] S1, determine the random sampling parameter N; and according to the random sampling parameter N, divide the pre-selected pairing sequence into two parts to obtain the first subsequence formed by the first N pairs and the second subsequence formed by the remaining pairs;

[0061] S2, based on the principle of the shortest possible path in the first subsequence, the Hungarian algorithm is used to pair and optimize all UAVs in the first subsequence to obtain the updated subsequence;

[0062] S3, determine the ideal path for each pairing in the updated subsequence, and add the updated subsequence to the second subsequence in descending order of the ideal paths to generate candidate pairing sequences;

[0063] S4, determine the random sampling parameter N, select the third subsequence formed by the first N pairs from the candidate pairing sequence, and determine the average path corresponding to the third subsequence based on the ideal path corresponding to each pair in the third subsequence;

[0064] S5, take the candidate pairing sequence as the new pre-selected pairing sequence, repeat steps S1 to S4 until the average path of the third subsequence no longer changes or the number of iterations reaches the maximum, then end the above operation and generate the optimized pairing sequence.

[0065] It should be noted that the random sampling parameter N that appears in each iteration or in the same iteration may be the same or different.

[0066] The random sampling parameter N is obtained by the following method: generating the random sampling parameter N based on the number of drones in the drone cluster corresponding to the pairing sequence, the pairing sequence parameters, and the dynamic adjustment parameters.

[0067] For example: the random sampling parameter N is set to a random, time-varying integer, i.e.

[0068] N = αβn (Equation 4);

[0069] Where n is the number of drones in the drone cluster corresponding to the pre-selected pairing sequence or candidate pairing sequence; β is the sequence parameter, which can generally be set to 0 < β ≤ 0.2; α is a dynamic adjustment parameter, 1 ≤ α ≤ 2, and α is randomly generated in each iteration step.

[0070] Because drone pairing is limited by the sequential pairing of start and target nodes, drones that are later in the pairing sequence are not optimal in their pairing choices, increasing the length of the ideal drone path. To address this, by performing local optimization on the pairing, the pairing order of drones can be optimized, thereby reducing the maximum path length among multiple drone paths, eliminating abnormal drone pairings in the pre-selected pairing sequence, and further reducing the average path and path standard deviation of the drone swarm. This achieves the minimization of the ideal path and balanced drone energy consumption, improving the efficiency of drone swarm management.

[0071] S203 specifically includes the following steps:

[0072] S1, based on the ideal path corresponding to each pair in the optimized pairing sequence, sort the optimized pairing sequence in descending order to generate an ordered pairing sequence;

[0073] S2, select the pair with the largest ideal path from the ordered pairing sequence as the target pair;

[0074] S3, perform drone node replacement operation for target pairing, and generate updated target pairing;

[0075] S4. If the ideal path of the updated target pair is less than that of the original target pair, then the updated target pair will replace the original target pair.

[0076] S5. Repeat steps S1, S2, S3, S4, and S5 until the total length of the ideal path corresponding to the optimized pairing sequence remains unchanged. Then, end the UAV node replacement operation in the target pairing and output the quasi-pairing sequence.

[0077] For example: Step 3.1: For any pair in the optimized pairing sequence: obtain the starting node corresponding to the original configuration and the target node corresponding to the target configuration in the pair; based on the coordinate information corresponding to the starting node and the coordinate information corresponding to the target node, determine the ideal path of the UAV corresponding to the pair; based on the ideal path corresponding to each pair, sort the optimized pairing sequence in descending order to generate an ordered pairing sequence;

[0078] Step 3.2: Select the pair with the largest ideal path from the ordered pairing sequence as the target pair, and calculate the ideal path length r from the starting node to all target nodes;

[0079] Step 3.2: According to the optimization rules for pairing the starting node and the target node, change the UAV pairing sequence; for example, try pairing the m-th UAV with other target nodes and calculate the new ideal path; if the ideal path of the updated target pairing is shorter than the original target pairing, then replace the original target pairing with the updated target pairing. The new pairings and their ideal paths are recorded below.

[0080] C m ={s m ,g m},C n ={s n ,g n}

[0081] If min{r n ,r m}<r mn ,r nm <max{r n ,r m}, then C m ={s m ,g n},C n ={s n ,g m}

[0082] Equation (5);

[0083] Among them, C m and C n This represents the pairing of the m-th and n-th drones, r m and r n Both represent the ideal path length of the original target pairing in Euclidean space; r mn ,r nm Both represent the ideal path length for updating the target pair.

[0084] Step 3.3: Repeat steps 3.1-3.2 until the total length of the ideal path corresponding to the optimized pairing sequence remains unchanged. Then, end the UAV node replacement operation in the target pairing and output the quasi-pairing sequence.

[0085] Therefore, by optimizing the pairing of individual drones, the pairing results in the original and target configurations can be optimized in a secondary manner, thereby eliminating the maximum ideal path length and balancing the ideal path length of drones. This reduces the standard deviation of the drone swarm, achieves balanced drone energy consumption, and improves the management efficiency of the drone swarm.

[0086] Traditional A* algorithms can only avoid static obstacles and cannot handle collisions between multiple UAVs. The improved A* algorithm, however, focuses on solving dynamic obstacle avoidance problems during multi-UAV flight. This embodiment introduces obstacle repulsion and conflict resolution strategies into the artificial potential field method, thereby solving the technical problems of static obstacle avoidance and real-time dynamic obstacle avoidance in large-scale UAV path planning. This enables effective path planning for transformations between different configurations, significantly improving the safety and efficiency of UAV swarms flying in complex environments.

[0087] like Figure 3The diagram shown is a flowchart illustrating the path planning process for generating unmanned vehicle configurations according to an embodiment of the present invention.

[0088] Perform path avoidance processing on each pair in the alignment sequence to generate a path plan for the unmanned machine, which includes at least the following steps:

[0089] S301, number the drones corresponding to the starting node of each pairing in the quasi-pairing sequence to obtain a number of numbered drones;

[0090] S302, for any current pairing in the quasi-pairing sequence: determine the current number of the UAV corresponding to the current pairing; control the current number UAV to receive path plans sent by other numbered UAVs in real time, and broadcast the path plans of the current number UAV in real time; obtain the current path node and target node corresponding to the current pairing; wherein, the current path node is used to indicate the starting node or an intermediate node located between the starting node and the target node;

[0091] S303, Based on the current path node and the target node, the next path node adjacent to the current path node is planned according to the static obstacle avoidance method to obtain the first path plan of the next path node;

[0092] S304, Based on the path planning information received by the UAV with the current number, the first path plan is detected;

[0093] S305, if the detection result indicates that there is a dynamic obstacle between the current path node and the next path node and the dynamic obstacle is not a path conflict, then the first path planning is determined to be invalid, and the next path node is replanned according to the dynamic obstacle avoidance method to generate the second path planning corresponding to the next path node; then S308 is executed.

[0094] S306, if the detection result indicates that there is a dynamic obstacle between the current path node and the next path node and the dynamic obstacle is a path conflict, then the next path node is replanned based on the dynamic programming method to generate the second path plan corresponding to the next path node; then S308 is executed.

[0095] S307, if the detection result indicates that there are no dynamic obstacles between the current path node and the next path node, then the next path node corresponding to the first path plan is taken as the quasi-path node.

[0096] S308, take the next path node corresponding to the second path plan as the quasi-path node;

[0097] S309, Repeat steps S301 to S308 until the current numbered UAV reaches the target node via different quasi-path nodes, and use the reached target node as a fixed obstacle to generate the path plan corresponding to the current pairing;

[0098] S310 generates path planning for unmanned structures based on the path planning corresponding to each pair in the quasi-pairing sequence.

[0099] For example, based on the current path node and the target node, a first path plan for the next path node adjacent to the current path node is obtained according to a static obstacle avoidance processing method; this includes: obtaining the attractive force of the target node on the current path node and the gravitational potential field corresponding to the attractive force; obtaining the repulsive force of the static obstacle on the current path node and the repulsive potential field corresponding to the repulsive force; determining the direction corresponding to the next path node adjacent to the current path node based on the attractive force and the repulsive force; determining the position information corresponding to the next path node adjacent to the current path node based on the gravitational potential field and the repulsive potential field; and determining the position information and direction corresponding to the next path node as the orientation information of the next path node, thereby obtaining the first path plan for the next path node.

[0100] It should be noted that the methods for handling static obstacle avoidance are the same as those for handling dynamic obstacle avoidance, and will not be repeated here.

[0101] For example: Step 4.1: Calculate the ideal path corresponding to each pair in the quasi-pairing sequence, assign numbers to the drones with different starting nodes in the quasi-pairing sequence; and control each numbered drone to broadcast the location information of the target node corresponding to its starting node to other numbered drones.

[0102] Step 4.2: For each numbered UAV, the resultant force exerted by the target node and obstacles on the UAV is calculated based on the improved potential field function, and path planning is performed for the next path node adjacent to the current path node for the numbered UAV. The attraction force of the target node on the current path node and the corresponding function definition of the gravitational potential field are shown below:

[0103]

[0104] Among them, F att (g) represents the attractiveness of the target node to the current path node; U att (g) represents the gravitational potential field corresponding to the attraction; μ is the proportional gain coefficient; r(p,g) is the distance between the target node and the current path node.

[0105] To avoid the goal being unreachable problem, repulsive forces and repulsive potential fields are set up. The repulsive force F exerted by static or dynamic obstacles on the current path node is... reo(a) and the repulsive potential field U reo The function expression for (a) is shown below:

[0106]

[0107] The repulsive force F exerted by static or dynamic obstacles on the numbered UAV reo1 and its vertical force F reo2 The specific function expression is as follows:

[0108]

[0109]

[0110] Where k is the proportional gain coefficient; r(a,g) is the distance from the coordinates of a dynamic obstacle (other numbered UAVs) or a static obstacle to the current path node. a It is the threshold of the repulsive force range of the obstacle (safety constraint range); It is the correction factor, r g =r(p,q), generally set n=2. This coefficient makes the repulsive and attractive forces on the drone reaching the target node zero, thus solving the problem of target unreachability. F reo1 The direction is from the obstacle to the numbered drone, F reo2 The direction is perpendicular to F reo1 direction.

[0111] The next path node for the numbered UAV is planned based on the gradient descent method. Specifically: based on F... att (g) and F reo (a) The resultant force determines the direction of the UAV in the next path node adjacent to the current path node; the total potential field is determined based on the gravitational potential field and the repulsive potential field; the position information of the next path node is determined based on the total potential field, the step size coefficient and the position information of the current path node.

[0112] The expression for the location information of the next path node is as follows:

[0113]

[0114] Where α represents the step size coefficient, and U(p) is the total potential field; p k This indicates the location information of the current path node.

[0115] Step 4.3: Based on the UAV kinematic model, continuously virtual sample the quasi-path nodes of the UAV, and consider the safety constraints of the UAV during flight by introducing obstacle repulsion force to replan the next path point of the UAV;

[0116] Step 4.4: Replan the path nodes for path conflicts that are difficult to avoid in the pre-planning of UAVs based on dynamic programming technology;

[0117] Step 4.5: For drones that have reached the target node, set them as fixed obstacles and consider the repulsive force of static obstacles in the path planning of subsequent tasks.

[0118] Step 4.6: Repeat steps 4.2-4.5 until all numbered drones have reached their respective target node positions.

[0119] Existing artificial potential field methods are local planning algorithms that can achieve obstacle avoidance in both dynamic and static scenarios. However, they are prone to getting trapped in local optima and path unreachability, leading to task planning failure. This embodiment introduces obstacle repulsion and conflict reduction strategies into the artificial potential field method, achieving not only global optimization of the planned path but also solving the path unreachability problem in existing methods. This further addresses the real-time dynamic obstacle avoidance issue in large-scale UAV path planning and improves the reliability of path planning during UAV configuration changes.

[0120] The following section will provide a detailed description of a large-scale unmanned structure transformation method based on a descending order strategy, using a specific application scenario as an example.

[0121] A method for large-scale unmanned vehicle configuration transformation based on a descending order strategy includes at least the following steps:

[0122] S1. Extract the outer surface contour points from the 3D model corresponding to the first unmanned vehicle configuration, and scale the extracted outer surface contour points to generate the original configuration; obtain the second unmanned vehicle configuration corresponding to the first unmanned vehicle configuration; extract the outer surface contour points from the 3D model corresponding to the second unmanned vehicle configuration, and scale the extracted outer surface contour points to generate the target configuration.

[0123] S2, Based on the position information of each UAV in the target configuration, determine the centroid of the target configuration; determine the node distance from each starting node in the original configuration to the centroid; based on the node distances corresponding to the starting nodes, sort the UAV cluster in the original configuration in descending order to generate the original configuration sequence; determine the node distance from each target node in the target configuration to the centroid; based on the node distances corresponding to the target nodes, sort the UAV cluster in the target configuration in descending order to generate the target configuration sequence.

[0124] S3. Select the first starting node from the original configuration sequence and the first target node from the target configuration sequence. If the node distance corresponding to the first starting node is greater than the node distance corresponding to the first target node, determine the path length between the first starting node and each target node in the target configuration sequence, and select the target node with the shortest path length from the target configuration as the pairing node of the first starting node. If the node distance corresponding to the first starting node is less than the node distance corresponding to the first target node, determine the path length between the first target node and each starting node in the original configuration sequence, and select the starting node with the shortest path length from the original configuration as the pairing node of the first target node. Remove the successfully paired nodes from the original configuration sequence and the target configuration sequence respectively, and generate the first update sequence corresponding to the original configuration sequence and the second update sequence corresponding to the target configuration sequence. Repeat the above steps in sequence until all UAVs in the original configuration sequence or the target configuration sequence are paired, generating several pairs.

[0125] S4. Determine the ideal path for each pair in a number of pairs, and sort the pairs in descending order of the ideal paths to obtain a pre-selected pairing sequence.

[0126] S5, determine the random sampling parameter N; and according to the random sampling parameter N, divide the pre-selected pairing sequence into two parts to obtain the first subsequence formed by the first N pairs and the second subsequence formed by the remaining pairs; based on the principle of the shortest ideal path in the first subsequence, optimize the pairing of all UAVs in the first subsequence according to the Hungarian algorithm to obtain the updated subsequence; determine the ideal path of each pair in the updated subsequence, and add the updated subsequence to the second subsequence in descending order of the ideal paths to generate a candidate pairing sequence; determine the random sampling parameter N, and select the third subsequence formed by the first N pairs from the candidate pairing sequence, and determine the average path corresponding to the third subsequence based on the ideal path corresponding to each pair in the third subsequence; use the candidate pairing sequence as the new pre-selected pairing sequence, and repeat the above steps until the average path of the third subsequence no longer changes, and then end the above operation to generate the optimized pairing sequence.

[0127] S6. Based on the ideal path corresponding to each pair in the optimized pairing sequence, sort the optimized pairing sequence in descending order to generate an ordered pairing sequence; select the pair with the largest ideal path from the ordered pairing sequence as the target pairing; perform a UAV node replacement operation on the target pairing to generate an updated target pairing; if the ideal path of the updated target pairing is smaller than that of the original target pairing, replace the original target pairing with the updated target pairing; repeat the above steps until the total length of the ideal path corresponding to the optimized pairing sequence remains unchanged, then end the UAV node replacement operation in the target pairing and output the quasi-pairing sequence.

[0128] S7, number the drones corresponding to the starting nodes of each pairing in the quasi-pairing sequence to obtain several numbered drones; for any current pairing in the quasi-pairing sequence: determine the current numbered drone corresponding to the current pairing; control the current numbered drone to receive path plans sent by other numbered drones in real time, and broadcast the path plans of the current numbered drone in real time; obtain the current path node and target node corresponding to the current pairing; wherein, the current path node is used to indicate the starting node or an intermediate node located between the starting node and the target node; based on the current path node and the target node, plan the next path node adjacent to the current path node according to the static obstacle avoidance processing method to obtain the first path plan of the next path node; if the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node and the dynamic obstacle is not a path conflict, then determine If the first path plan fails, the next path node is replanned according to the dynamic obstacle avoidance method, generating a second path plan corresponding to the next path node. If the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node, and the dynamic obstacle is a path conflict, the next path node is replanned based on the dynamic planning method, generating a second path plan corresponding to the next path node. The next path node corresponding to the second path plan is taken as a quasi-path node. If the received path plan indicates that there is no dynamic obstacle between the current path node and the next path node, the next path node corresponding to the first path plan is taken as a quasi-path node. The above steps are repeated until the UAV with the current number reaches the target node through different quasi-path nodes, and the reached target node is taken as a fixed obstacle, generating the path plan corresponding to the current pair.

[0129] S8 generates path planning for unmanned structures based on the path planning corresponding to each pair in the quasi-pairing sequence.

[0130] This embodiment presents a configuration transformation method for UAV swarms based on a descending order strategy. This method effectively reduces the path length of individual UAVs and the swarm as a whole, thereby achieving balanced resource consumption within the swarm. Furthermore, this embodiment combines local and global optimization when pairing UAVs, enabling the optimization problem of UAV pairing and combination with relatively low computational cost. This reduces computational complexity while minimizing the ideal path of each pair and achieving energy balance, thus improving the management efficiency of large-scale UAV swarms.

[0131] like Figure 4 The diagram shown is a structural schematic of a large-scale unmanned structure transformation device based on a descending order strategy, according to an embodiment of the present invention.

[0132] A large-scale unmanned aerial vehicle (UAV) configuration transformation device based on a descending order strategy, the device 400 includes: a first generation module 401, used to arrange the UAV cluster in the original configuration in descending order based on the node distance between each starting node and a reference point in the original configuration, generating an original configuration sequence; a second generation module 402, used to arrange the UAV cluster in the target configuration in descending order based on the node distance between each target node and a reference point in the target configuration corresponding to the original configuration, generating a target configuration sequence; wherein, the starting node is used to indicate the azimuth information of the UAV in the original configuration; the target ... The marker node is used to indicate the orientation information of the UAV in the target configuration; the global pairing module 403 is used to perform global pairing of all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence in sequence to generate several pairs; the optimization processing module 404 is used to sort the several pairs in descending order according to the ideal path corresponding to the pair, and perform pairing optimization processing on the obtained pre-selected pairing sequence to obtain a quasi-pairing sequence; the path obstacle avoidance processing module 405 is used to perform path obstacle avoidance processing on each pair in the quasi-pairing sequence to generate the path planning of the UAV configuration.

[0133] In a preferred embodiment of this example, the global pairing module includes: a selection unit, configured to select a first starting node from the original configuration sequence and a first target node from the target configuration sequence; a first determination unit, configured to determine the path length between the first starting node and each target node in the target configuration sequence if the node distance corresponding to the first starting node is greater than the node distance corresponding to the first target node, and select the target node with the shortest path length from the target configuration as the pairing node of the first starting node; a second determination unit, configured to determine the path length between the first target node and each starting node in the original configuration sequence if the node distance corresponding to the first starting node is less than the node distance corresponding to the first target node, and select the starting node with the shortest path length from the original configuration as the pairing node of the first target node; a removal unit, configured to remove successfully paired nodes from the original configuration sequence and the target configuration sequence respectively, generating a first update sequence corresponding to the original configuration sequence and a second update sequence corresponding to the target configuration sequence; and a looping unit, configured to repeat the above steps sequentially until all UAVs in the original configuration sequence or the target configuration sequence have completed pairing, generating several pairings.

[0134] In a preferred embodiment of this example, the optimization processing module includes: a determination unit, configured to determine the ideal path corresponding to each of the plurality of pairings, and sort the plurality of pairings in descending order of the ideal paths to obtain a pre-selected pairing sequence; a local optimization unit, configured to select the first N pairings from the pre-selected pairing sequence by random sampling of nodes, and perform local optimization processing on the subsequences formed by the first N pairings to generate an optimized pairing sequence; and a single-unit optimization unit, configured to perform single-unit optimization processing on the optimized pairing sequence to obtain a quasi-pairing sequence.

[0135] In a preferred embodiment of this example, the local optimization unit includes: a partitioning subunit, used to determine a random sampling parameter N; and according to the random sampling parameter N, to divide the pre-selected pairing sequence into two parts, obtaining a first sub-sequence formed by the first N pairs and a second sub-sequence formed by the remaining pairs; a pairing optimization subunit, used to optimize the pairing of all UAVs in the first sub-sequence based on the principle of the shortest ideal path in the first sub-sequence and according to the Hungarian algorithm, to obtain an updated sub-sequence; a first generation subunit, used to determine the ideal path of each pair in the updated sub-sequence, and add the updated sub-sequence to the second sub-sequence in descending order of the ideal paths, generating a candidate pairing sequence; a determining subunit, used to determine the random sampling parameter N, and select a third sub-sequence formed by the first N pairs from the candidate pairing sequence, and determine the average path corresponding to the third sub-sequence based on the ideal path corresponding to each pair in the third sub-sequence; and a second generation subunit, used to take the candidate pairing sequence as a new pre-selected pairing sequence, and repeat the above steps sequentially until the average path of the third sub-sequence no longer changes, thus generating an optimized pairing sequence.

[0136] In a preferred embodiment of this example, the single-unit optimization unit includes: a descending order sorting subunit, used to sort the optimized pairing sequence in descending order based on the ideal path corresponding to each pair in the optimized pairing sequence to generate an ordered pairing sequence; a selection subunit, used to select the pair with the largest ideal path from the ordered pairing sequence as the target pairing; a replacement operation subunit, used to perform a UAV node replacement operation on the target pairing to generate an updated target pairing; an update subunit, used to replace the original target pairing with the updated target pairing if the ideal path of the updated target pairing is smaller than that of the original target pairing; and a looping subunit, used to repeat the above steps sequentially until the total length of the ideal path corresponding to the optimized pairing sequence remains unchanged, then the UAV node replacement operation in the target pairing ends and a quasi-pairing sequence is output.

[0137] In a preferred embodiment of this example, the path obstacle avoidance processing module includes: a numbering unit, used to number the UAVs corresponding to the starting node of each pairing in the quasi-pairing sequence, obtaining a plurality of numbered UAVs; a determination unit, used for any current pairing in the quasi-pairing sequence: determining the current numbered UAV corresponding to the current pairing; controlling the current numbered UAV to receive path plans sent by other numbered UAVs in real time, and broadcasting the path plans of the current numbered UAV in real time; obtaining the current path node and target node corresponding to the current pairing; wherein, the current path node is used to indicate the starting node or an intermediate node located between the starting node and the target node; a first planning unit, used to plan the next path node adjacent to the current path node based on the current path node and the target node, according to a static obstacle avoidance processing method, to obtain a first path plan for the next path node; a second planning unit, used to determine the first path if the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node and the dynamic obstacle is not a path conflict. If the current path plan is invalid, the next path node is replanned according to the dynamic obstacle avoidance method to generate a third path plan corresponding to the next path node; if the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node and the dynamic obstacle is a path conflict, the next path node is replanned based on the dynamic planning method to generate a second path plan corresponding to the next path node; a first determining unit is used to take the next path node corresponding to the second path plan as a quasi-path node; a second determining unit is used to take the next path node corresponding to the first path plan as a quasi-path node if the received path plan indicates that there is no dynamic obstacle between the current path node and the next path node; a loop repeating unit is used to repeat the above steps sequentially until the UAV with the current number reaches the target node via different quasi-path nodes, and the reached target node is taken as a fixed obstacle to generate the path plan corresponding to the current pair; a generation unit is used to generate a path plan for the UAV type based on the path plan corresponding to each pair in the quasi-pairing sequence.

[0138] In a preferred embodiment of this example, the first planning unit includes: a first acquisition subunit, used to acquire the attraction of the target node to the current path node and the gravitational potential field corresponding to the attraction; a second acquisition subunit, used to acquire the repulsion of a static obstacle to the current path node and the repulsion potential field corresponding to the repulsion; a first determination subunit, used to determine the direction corresponding to the next path node adjacent to the current path node based on the attraction and the repulsion; a second determination subunit, used to determine the position information corresponding to the next path node adjacent to the current path node based on the gravitational potential field and the repulsion potential field; and a third determination subunit, used to determine the position information and direction corresponding to the next path node as the orientation information of the next path node, thereby obtaining the first path planning for the next path node.

[0139] The device further includes: a first determining module, configured to determine the centroid of the target configuration based on the position information of each UAV in the target configuration; a second determining module, configured to determine the node distance from each starting node in the original configuration to the centroid; a first descending order arranging module, configured to arrange the UAV cluster in the original configuration in descending order based on the node distance corresponding to the starting node, generating an original configuration sequence; a third determining module, configured to determine the node distance from each target node in the target configuration to the centroid; and a second descending order arranging module, configured to arrange the UAV cluster in the target configuration in descending order based on the node distance corresponding to the target node, generating a target configuration sequence.

[0140] In a preferred embodiment of this invention, the device further includes: a first processing module, configured to extract outer surface contour points from the three-dimensional model corresponding to the first unmanned vehicle configuration, and to scale the extracted outer surface contour points to generate an original configuration; an acquisition module, configured to acquire a second unmanned vehicle configuration corresponding to the first unmanned vehicle configuration; and a second processing module, configured to extract outer surface contour points from the three-dimensional model corresponding to the second unmanned vehicle configuration, and to scale the extracted outer surface contour points to generate a target configuration.

[0141] The aforementioned device can execute a large-scale unmanned machine configuration transformation method based on a descending order strategy provided in an embodiment of the present invention, and possesses the corresponding functional modules and beneficial effects for executing such a method. Technical details not described in detail in this embodiment can be found in the large-scale unmanned machine configuration transformation method based on a descending order strategy provided in an embodiment of the present invention.

[0142] The present invention also provides an electronic device, comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the large-scale unmanned structure transformation method based on a descending order strategy described in the present invention.

[0143] In addition to the methods and apparatus described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of this application described in the "Exemplary Methods" section above.

[0144] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0145] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps of the methods according to the following embodiments of this application as described in the "Exemplary Methods" section above.

[0146] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0147] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0148] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0149] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0150] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0151] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for large-scale unmanned vehicle configuration transformation based on a descending order strategy, characterized in that, include: Based on the node distances from each starting node to the reference point in the original configuration, the UAV cluster in the original configuration is sorted in descending order to generate the original configuration sequence. Based on the node distance from each target node to the reference point in the target configuration corresponding to the original configuration, the UAV cluster in the target configuration is sorted in descending order to generate a target configuration sequence; wherein, the starting node is used to indicate the azimuth information of the UAV in the original configuration; the target node is used to indicate the azimuth information of the UAV in the target configuration; The starting nodes in the original configuration sequence and the target nodes in the target configuration sequence are globally paired sequentially to generate several pairs; The pairings are arranged in descending order according to the ideal path corresponding to each pairing, and the pre-selected pairing sequence obtained by the descending order is subjected to pairing optimization processing to obtain a quasi-pairing sequence. For each pairing in the quasi-pairing sequence, path obstacle avoidance processing is performed to generate a path plan for the unmanned mechanism. The step of arranging the several pairs in descending order according to the ideal paths corresponding to the pairs, and performing pair optimization processing on the pre-selected pairing sequence obtained by descending order to obtain a quasi-pairing sequence includes: determining the ideal path corresponding to each pair in the several pairs, and sorting the several pairs according to the descending order of the ideal paths to obtain a pre-selected pairing sequence; selecting the first N pairs from the pre-selected pairing sequence by random sampling of nodes, and performing local optimization processing on the subsequences formed by the first N pairs to generate an optimized pairing sequence; and performing individual optimization processing on the optimized pairing sequence to obtain a quasi-pairing sequence.

2. The method according to claim 1, characterized in that, The process involves sequentially performing global pairing of all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence to generate several pairs. include: Select the first starting node from the original configuration sequence, and select the first target node from the target configuration sequence; If the node distance corresponding to the first starting node is greater than the node distance corresponding to the first target node, then the path length between the first starting node and each target node in the target configuration sequence is determined, and the target node with the shortest path length is selected from the target configuration as the paired node of the first starting node. If the node distance corresponding to the first starting node is less than the node distance corresponding to the first target node, then the path length between the first target node and each starting node in the original configuration sequence is determined, and the starting node with the shortest path length is selected from the original configuration as the paired node of the first target node. Remove successfully paired nodes from the original configuration sequence and the target configuration sequence respectively to generate a first update sequence corresponding to the original configuration sequence and a second update sequence corresponding to the target configuration sequence; Repeat the above steps in sequence until all UAVs in the original configuration sequence or the target configuration sequence are paired, generating several pairs.

3. The method according to claim 1, characterized in that, The step of selecting the first N pairs from the pre-selected pairing sequence by random sampling of nodes, and performing local optimization processing on the subsequences formed by the first N pairs to generate an optimized pairing sequence includes: Determine the random sampling parameter N; and according to the random sampling parameter N, divide the preselected pairing sequence into two parts to obtain a first subsequence formed by the first N pairs and a second subsequence formed by the remaining pairs; Based on the principle of finding the shortest path in the first subsequence, the Hungarian algorithm is used to pair and optimize all UAVs in the first subsequence to obtain the updated subsequence. Determine the ideal path for each pairing in the updated subsequence, and add the updated subsequence to the second subsequence in descending order of the ideal paths to generate candidate pairing sequences; Determine the random sampling parameter N, select the third subsequence formed by the first N pairs from the candidate pairing sequence, and determine the average path corresponding to the third subsequence based on the ideal path corresponding to each pair in the third subsequence; The candidate pairing sequence is used as a new pre-selected pairing sequence. The above steps are repeated until the average path of the third subsequence no longer changes, and the above operation is ended to generate the optimized pairing sequence.

4. The method according to claim 1, characterized in that, The step of performing single-unit optimization on the optimized paired sequence to obtain a quasi-paired sequence includes: Based on the ideal path corresponding to each pair in the optimized pairing sequence, the optimized pairing sequence is sorted in descending order to generate an ordered pairing sequence. Select the pair with the largest ideal path from the ordered pairing sequence as the target pair; Perform a drone node replacement operation on the target pairing to generate an updated target pairing; If the ideal path of the updated target pair is shorter than the original target pair, then the updated target pair replaces the original target pair. Repeat the above steps until the total length of the ideal path corresponding to the optimized pairing sequence remains unchanged. Then, end the UAV node replacement operation in the target pairing and output the quasi-pairing sequence.

5. The method according to claim 1, characterized in that, The step of performing path obstacle avoidance processing on each pair in the quasi-pairing sequence to generate a path plan for the unmanned vehicle type includes: The drones corresponding to the starting nodes of each pairing in the quasi-pairing sequence are numbered to obtain a number of numbered drones; For any current pairing in the quasi-pairing sequence: determine the current number of the UAV corresponding to the current pairing; control the current number UAV to receive path plans sent by other numbered UAVs in real time, and broadcast the path plans of the current number UAV in real time; obtain the current path node and target node corresponding to the current pairing; wherein, the current path node is used to indicate the starting node or an intermediate node located between the starting node and the target node; Based on the current path node and the target node, the next path node adjacent to the current path node is planned according to the static obstacle avoidance method to obtain the first path plan of the next path node. If the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node, and the dynamic obstacle is not a path conflict, then the first path plan is determined to be invalid, and the next path node is replanned according to the dynamic obstacle avoidance method to generate a second path plan corresponding to the next path node; if the received path plan indicates that there is a dynamic obstacle between the current path node and the next path node, and the dynamic obstacle is a path conflict, then the next path node is replanned based on the dynamic planning method to generate a second path plan corresponding to the next path node. The next path node corresponding to the second path plan is taken as the quasi-path node; If the received path plan indicates that there are no dynamic obstacles between the current path node and the next path node, then the next path node corresponding to the first path plan is taken as the quasi-path node. Repeat the above steps in sequence until the UAV with the current number reaches the target node via different quasi-path nodes, and use the reached target node as a fixed obstacle to generate the path plan corresponding to the current pairing; Based on the path planning corresponding to each pair in the quasi-pairing sequence, a path planning for the unmanned mechanism is generated.

6. The method according to claim 5, characterized in that, The step of planning the next path node adjacent to the current path node based on the current path node and the target node, according to a static obstacle avoidance method, to obtain a first path plan for the next path node; includes: Obtain the attraction between the target node and the current path node, as well as the gravitational potential field corresponding to the attraction. Obtain the repulsive force exerted by static obstacles on the current path node and the repulsive potential field corresponding to the repulsive force; Based on the attractive and repulsive forces, determine the direction corresponding to the next path node adjacent to the current path node; Based on the gravitational potential field and the repulsive potential field, determine the position information of the next path node adjacent to the current path node; The location information and direction corresponding to the next path node are determined as the orientation information of the next path node, thereby obtaining the first path plan for the next path node.

7. The method according to claim 1, characterized in that, Based on the position information of each UAV in the target configuration, the centroid of the target configuration is determined; Determine the node distance from each starting node to the centroid in the original configuration; Based on the node distances corresponding to the starting node, the drone clusters in the original configuration are sorted in descending order to generate the original configuration sequence. Determine the node distance from each target node to the centroid in the target configuration; Based on the node distances corresponding to the target nodes, the UAV clusters in the target configuration are sorted in descending order to generate a target configuration sequence.

8. The method according to claim 1, characterized in that, Also includes: Extract the outer surface contour points from the 3D model corresponding to the first unmanned structure, and scale the extracted outer surface contour points to generate the original structure. Obtain the second unmanned mechanism type corresponding to the first unmanned mechanism type; Extract the outer surface contour points from the three-dimensional model corresponding to the second unmanned structure, and scale the extracted outer surface contour points to generate the target structure.

9. A large-scale unmanned structure transformation device based on a descending order strategy, characterized in that, include: The first generation module is used to sort the UAV cluster in the original configuration in descending order based on the node distance between each starting node and the reference point in the original configuration, and generate the original configuration sequence. The second generation module is used to sort the UAV cluster in the target configuration in descending order based on the node distance between each target node in the target configuration corresponding to the original configuration and the reference point, thereby generating a target configuration sequence; wherein, the starting node is used to indicate the azimuth information of the UAV in the original configuration; and the target node is used to indicate the azimuth information of the UAV in the target configuration; The global pairing module is used to perform global pairing of all starting nodes in the original configuration sequence and all target nodes in the target configuration sequence in sequence to generate several pairs. An optimization processing module is used to sort the several pairs in descending order according to the ideal paths corresponding to the pairs, and to perform pair optimization processing on the obtained pre-selected pairing sequence to obtain a quasi-pairing sequence. The optimization processing module includes: a determination unit, used to determine the ideal path corresponding to each pair in the several pairs, and to sort the several pairs in descending order according to the ideal paths to obtain a pre-selected pairing sequence; a local optimization unit, used to select the first N pairs from the pre-selected pairing sequence by random sampling of nodes, and to perform local optimization processing on the subsequences formed by the first N pairs to generate an optimized pairing sequence; and a single-unit optimization unit, used to perform single-unit optimization processing on the optimized pairing sequence to obtain a quasi-pairing sequence. The path obstacle avoidance processing module is used to perform path obstacle avoidance processing on each pair in the quasi-pairing sequence to generate path planning for the unmanned mechanism.

10. A computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method as claimed in any one of claims 1 to 8.

Images