Flight control method for scheduling massive unmanned aerial vehicle cluster work

Abstract

The invention belongs to the technical field of unmanned aerial vehicles, and particularly discloses a flight control method for scheduling a large number of unmanned aerial vehicle cluster operations, and the method comprises the following steps: S1, finding out an optimal path through an unmanned aerial vehicle energy-saving path planning obstacle avoidance algorithm; s2, selecting the most energy-saving path for flight through a cost function; s3, establishing a cluster unmanned aerial vehicle space protection model; s4, formulating an anti-collision strategy and an anti-collision algorithm to avoid potential collision conflicts; s5, formulating an unmanned aerial vehicle collision autonomous flight obstacle avoidance strategy to guarantee safe flight of the unmanned aerial vehicle; according to the invention, a plurality of key technical fields of an unmanned aerial vehicle cluster cooperation technology, an intelligent flight control technology, a communication technology, a simulation and test technology and an artificial intelligence and machine learning technology are integrated, and efficient cooperation operation and intelligent scheduling of the unmanned aerial vehicle cluster are realized, so that the unmanned aerial vehicle cluster can better adapt to complex environments and task requirements; and the flight efficiency of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicles, and in particular relates to a flight control method for scheduling a large number of unmanned aerial vehicle cluster operations. Background Art

[0002] With the rapid development of neural networks, machine learning, deep learning and other methods, the control of UAV swarms using machine learning methods has developed rapidly in recent years. Although UAV swarm control methods have made significant progress in many aspects, there are still some limitations; the flight energy saving, flight collision, and flight avoidance of swarm UAVs need to be optimized and improved. UAV energy is a bottleneck. Most UAVs rely on lithium batteries for power supply, and the energy density of lithium batteries is limited, which directly limits the flight time of UAVs. It is necessary to develop effective methods for energy-saving flight. The current technology considers flight route planning, but lacks consideration in terms of flight energy consumption. To this end, we propose a flight control method for scheduling a large number of UAV swarm operations to solve the problems existing in the existing technology, so that it can improve the flight efficiency of UAVs. Summary of the invention

[0003] The purpose of the present invention is to provide a flight control method for scheduling a large number of drone swarm operations, which can improve the performance, stability, and task execution efficiency of the drone swarm to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] A flight control method for scheduling a large number of unmanned aerial vehicle swarm operations comprises the following steps:

[0006] S1, find the optimal path through the drone energy-saving path planning and obstacle avoidance algorithm;

[0007] S2, select the most energy-efficient flight path through the cost function;

[0008] S3, establish a cluster UAV space protection model;

[0009] S4. Develop anti-collision strategies and anti-collision algorithms to avoid potential collision conflicts;

[0010] S5. Develop autonomous flight obstacle avoidance strategies for UAV collisions to ensure safe flight of UAVs.

[0011] Preferably, in step S1, the process of the drone energy-saving path planning and obstacle avoidance algorithm to find the optimal path is as follows:

[0012] S11. Initialize the information of the UAV cluster. The UAV cluster information includes terrain and landform, obstacle positions, climate conditions, and flight missions. The flight missions include initial positions, target routes, and flight times, which are collected through a sensor group. The sensor group includes radar, lidar, and cameras.

[0013] S12. Generate a pre-planned route. Obtain global information based on the UWB positioning system, and then initially generate a pre-planned path that can connect the starting point and the target point of the UAV according to the position of the target point.

[0014] S13. Generate candidate paths. Generate a series of candidate paths by sampling the global path, and then use the graph search algorithm to search for the sampling points to obtain the path with the shortest distance or the minimum cost as the candidate path.

[0015] Preferably, in step S12, the pre-planned path generation rule is as follows: Connect the starting point and the target point; Determine whether there is an obstacle intersecting the obtained connection line. If so, select the obstacle closest to the starting point that intersects the connection; Select a new path point from the nearest obstacle and set it as the starting point.

[0016] Preferably, in step S13, the process of generating candidate paths is as follows:

[0017] A1. Create an auxiliary array D to store the shortest path lengths from the starting point to each other vertex.

[0018] A2. Set the shortest path length of the starting point to 0, and the shortest path lengths of other vertices to infinity.

[0019] A3. Create a set S to store the visited vertices. Initially, S only contains the starting point.

[0020] A4. Select the vertex u closest to the starting point from the unvisited vertices, that is, the vertex with the smallest D[u], and add it to the set S.

[0021] A5. For each neighbor v of vertex u, if the path through u to v is shorter than the currently recorded path from the starting point to v, that is, D[u] + weight(u, v) < D[v], then update D[v] to D[u] + weight(u, v).

[0022] A6. Repeat the above steps A4 - A5 until all vertices are visited or a certain termination condition is reached.

[0023] Preferably, in step S2, the cost function is J(i) = ∑(w saf f saf (i) + w smo f smo(i)+w ene f ene (i)), where J(i) is the total cost of path i, i is the candidate path index, and f saf is the safety cost function, expressed as the inverse of the minimum distance from the candidate path to the obstacle edge, f smo (i) is the smoothing cost function, represented by the maximum curvature in the candidate path, f ene It is the energy saving cost function, which is expressed by the effective work when the candidate path is executed.

[0024] Preferably, in step S3, the cluster UAV space protection model is a two-layer space protection model, which is divided into an alarm area and a warning area. The warning area is a prototype area with the center of mass of the UAV as the center and the minimum safety distance Q as the radius; the warning area is a prototype area with the center of mass of the UAV as the center and 2Q as the radius.

[0025] Preferably, the swarm drone space protection model receives and processes the flight data of the drone swarm in real time, and determines whether there is a potential collision conflict according to a preset rule or algorithm. When constructing the model, it is necessary to consider the influence of the drone's motion characteristics, flight environment and communication factors on conflict detection. The model rules are as follows:

[0026] S31. When an obstacle appears in the warning area, it is considered as no collision conflict, the detection model is warned, and the dynamic avoidance warning mechanism between drones is activated;

[0027] S32, when the obstacle appears in the warning area but outside the alarm area, determining the relationship between the speed vector of the obstacle and the speed vector of the drone;

[0028] S33: When an obstacle appears in the warning area, it is considered a collision.

[0029] S34: When the distance between the obstacle and the drone is less than one diameter of the drone, it is considered a collision.

[0030] Preferably, in step S4, the anti-collision strategy includes adjusting the parameters of the flight trajectory, speed or altitude of the drone to avoid potential collision conflicts; the anti-collision algorithm adopts the dynamic frame time slot Aloha algorithm to estimate the tag backlog according to the number of collision time slots by dividing the time slots and dynamically adjusting the frame length.

[0031] Preferably, the process of estimating the label backlog by the dynamic frame time slot Aloha algorithm is as follows:

[0032] C1, query time slot,

[0033] C2, calculate E, S, C at the i-th time slot to get the estimated value n;

[0034] C3. Determine whether n is within the optimal range. If so, select the next time slot and repeat C2. If not, change the Q value according to the estimated backlog and return to C2.

[0035] Preferably, in step S5, the UAV collision autonomous flight obstacle avoidance strategy adopts a cluster avoidance strategy of secondary decision, and calculates the minimum turning radius of the UAV according to the maximum acceleration of the UAV and the current flight speed, that is, the maximum turning angle of the UAV. Between the two maximum turning angles on the left and right of the UAV's forward flight direction, with an iteration time interval of 0.1s and an angle interval of 1 degree, the UAV can generate 13 path points at the next moment, and a secondary decision is made on the 13 path points generated by the UAV at the next moment to ensure the safe flight of the UAV, wherein the 13 path points include 6 left-turn path points, 6 right-turn path points and 1 straight-line flight path point.

[0036] The flight control method for scheduling a large number of UAV swarm operations proposed in the present invention has the following advantages compared with the prior art:

[0037] 1. The present invention finds the optimal path through the unmanned aerial vehicle energy-saving path planning and obstacle avoidance algorithm, selects the most energy-saving path flight through the cost function, establishes a cluster unmanned aerial vehicle space protection model, formulates anti-collision strategies and anti-collision algorithms and unmanned aerial vehicle collision autonomous flight obstacle avoidance strategies, integrates multiple key technical fields such as unmanned aerial vehicle cluster collaboration technology, intelligent flight control technology, communication technology, simulation and testing technology, and artificial intelligence and machine learning technology, and realizes efficient collaborative operation and intelligent scheduling of unmanned aerial vehicle clusters through advanced algorithms and communication technologies, including information sharing, task allocation, path planning, obstacle avoidance flight and other functions. This method improves the autonomous flight capability and intelligent decision-making level of unmanned aerial vehicles, enables unmanned aerial vehicle clusters to better adapt to complex environments and mission requirements, and improves the flight efficiency of unmanned aerial vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flow chart of the present invention;

[0039] Figure 2 A flowchart of the optimal path search for the energy-saving path planning and obstacle avoidance algorithm of the UAV of the present invention;

[0040] Figure 3 It is a schematic diagram of the obstacle avoidance circuit of the present invention;

[0041] Figure 4 This is a schematic diagram of the drone protection area model of the present invention;

[0042] Figure 5 The flow chart of the dynamic frame time slot Aloha algorithm of the present invention;

[0043] Figure 6 It is a schematic diagram of the set model of the secondary decision boundary obstacle avoidance mechanism of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0045] The present invention provides Figure 1-6 A flight control method for scheduling a large number of UAV swarm operations is shown, comprising the following steps:

[0046] S1, find the optimal path through the drone energy-saving path planning and obstacle avoidance algorithm;

[0047] The energy-saving path planning and obstacle avoidance algorithm for drones increases the continuity of the path by generating a pre-planned path, and then generates multiple candidate paths by adjusting the potential field coefficients. It also calculates the energy consumption of each candidate path while satisfying the physical constraints of the drone and ensuring the safety of the path, thereby selecting an optimal path.

[0048] like Figure 2 As shown, the process of the UAV energy-saving path planning and obstacle avoidance algorithm to find the optimal path is as follows:

[0049] S11, initializing the drone cluster information, the drone cluster information includes topography, obstacle locations, climate conditions and flight missions, wherein the flight missions include initialization locations, target routes and flight time, which are collected through a sensor group, which includes radar, lidar and camera;

[0050] S12, pre-planned route generation, based on the global information obtained by the UWB positioning system, and then based on the position of the target point, a pre-planned path that can connect the starting point and the target point of the drone is initially generated; the pre-planned path generation rule is: connect the starting point and the target point; determine whether there is an obstacle intersecting with the obtained connection line, if so, select the obstacle closest to the starting point that intersects with the connection; select a new path point from the nearest obstacle and set it as the starting point, such as Figure 3 Shown is a schematic diagram of the obstacle avoidance route;

[0051] S13. Candidate path generation: By sampling the global path, a series of candidate paths are generated, and then the graph search algorithm is used to search for the sampling points. The graph search algorithm includes Dijkstra's algorithm and A* algorithm to obtain the path with the shortest distance or the minimum cost as the candidate path;

[0052] The process of candidate path generation is as follows:

[0053] A1. Create an auxiliary array D to store the shortest path lengths from the starting point to each other vertex;

[0054] A2. Set the shortest path length of the starting point to 0, and the shortest path lengths of other vertices to infinity. It is also possible to set the shortest path lengths of other vertices to a sufficiently large number to indicate that they have not been visited;

[0055] A3. Create a set S to store the visited vertices. Initially, S only contains the starting point;

[0056] A4. Select the vertex u closest to the starting point from the unvisited vertices, that is, the vertex with the smallest D[u], and add it to the set S;

[0057] A5. For each neighbor v of vertex u, if the path through u to v is shorter than the currently recorded path from the starting point to v, that is, D[u] + weight(u, v) < D[v], then update D[v] to D[u] + weight(u, v);

[0058] A6. Repeat the above steps A4 - A5 until all vertices have been visited or a certain termination condition is reached, such as finding the shortest path to a specific target vertex.

[0059] S2. Select the most energy - efficient path for flight through the cost function;

[0060] The cost function is J(i) = ∑(w saf f saf (i) + w smo f smo (i) + w ene f ene (i)), where J(i) is the total cost of path i, i is the candidate path label, f saf is the safety cost function, represented by the reciprocal of the minimum distance from the candidate path to the obstacle edge, f smo (i) is the smoothness cost function, represented by the maximum curvature in the candidate path, f ene is the energy - saving cost function, represented by the effective work during the execution of the candidate path.

[0061] S3. Establish a spatial protection model for cluster drones;

[0062] The swarm UAV space protection model is a two-layer space protection model, which is divided into an alarm zone and a warning zone. The warning zone is a prototype zone with the center of mass of the UAV as the center and the minimum safety distance Q as the radius; the warning zone is a prototype zone with the center of mass of the UAV as the center and 2Q as the radius. Figure 4 Shown is a model of a drone protection area.

[0063] According to the selected data source and monitoring technology, a cluster UAV collision conflict detection model is constructed. The cluster UAV space protection model receives and processes the flight data of the UAV cluster in real time, and determines whether there is a potential collision conflict according to the preset rules or algorithms. When constructing the model, it is necessary to consider the influence of the UAV's motion characteristics, flight environment and communication factors on conflict detection. The model rules are as follows:

[0064] S31. When an obstacle appears in the warning area, it is considered as no collision conflict, the detection model is warned, and the dynamic avoidance warning mechanism between drones is activated;

[0065] S32, when the obstacle appears in the warning area but outside the alarm area, determining the relationship between the speed vector of the obstacle and the speed vector of the drone;

[0066] S33: When an obstacle appears in the warning area, it is considered a collision.

[0067] S34: When the distance between the obstacle and the drone is less than one diameter of the drone, it is considered a collision.

[0068] S4. Develop anti-collision strategies and anti-collision algorithms to avoid potential collision conflicts;

[0069] The anti-collision strategy includes adjusting the parameters of the UAV's flight trajectory, speed or altitude to avoid potential collision conflicts; the anti-collision algorithm adopts the dynamic frame time slot Aloha algorithm to significantly improve the recognition efficiency by dividing the time slots and dynamically adjusting the frame length. This method is a simple and effective way to reduce the computational complexity of estimating the number of collision tags. Based on the Schoute method, a feasible estimation method is proposed to estimate the tag backlog according to the number of collision time slots.

[0070] like Figure 5 As shown, the process of estimating the label backlog by the dynamic frame time slot Aloha algorithm is as follows:

[0071] C1, query time slot,

[0072] C2, calculate E, S, C at the i-th time slot to get the estimated value n;

[0073] C3. Determine whether n is within the optimal range. If so, select the next time slot and repeat C2. If not, change the Q value according to the estimated backlog and return to C2.

[0074] S5. Develop autonomous flight obstacle avoidance strategies for UAVs to ensure safe flight of UAVs;

[0075] The autonomous flight obstacle avoidance strategy for UAV collision adopts a cluster avoidance strategy with secondary decision-making, such as Figure 6 The figure shows a quadratic decision boundary obstacle avoidance mechanism ensemble model. The minimum turning radius of the drone, that is, the maximum turning angle of the drone, is calculated based on the maximum acceleration of the drone and the current flight speed. Between the two maximum turning angles on the left and right of the drone's forward flight direction, with an iteration time interval of 0.1s and an angle interval of 1 degree, the drone can generate 13 path points at the next moment. A secondary decision is made on the 13 path points generated by the drone at the next moment to ensure the safe flight of the drone. The 13 path points include 6 left-turn path points, 6 right-turn path points and 1 straight-line flight path point.

[0076] The optimal path is found through the UAV energy-saving path planning and obstacle avoidance algorithm, the most energy-saving path is selected through the cost function, a cluster UAV space protection model is established, and anti-collision strategies and anti-collision algorithms and UAV collision autonomous flight obstacle avoidance strategies are formulated. It integrates multiple key technical fields such as UAV cluster collaboration technology, intelligent flight control technology, communication technology, simulation and testing technology, and artificial intelligence and machine learning technology. Through advanced algorithms and communication technologies, efficient collaborative operation and intelligent scheduling of UAV clusters are realized, including information sharing, task allocation, path planning, obstacle avoidance flight and other functions. This method improves the autonomous flight capability and intelligent decision-making level of UAVs, enables UAV clusters to better adapt to complex environments and mission requirements, and improves the flight efficiency of UAVs.

[0077] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A flight control method for scheduling a large number of UAV swarm operations, characterized in that: It includes the following steps: S1. Find the optimal path through the energy-saving path planning and obstacle avoidance algorithm for drones; S2. Select the most energy-saving path for flight through the cost function; S3. Establish a spatial protection model for cluster drones; S4. Develop a collision avoidance strategy and algorithm to avoid potential collision conflicts; S5. Develop an autonomous flight obstacle avoidance strategy for drone collisions to ensure the safe flight of drones.

2. A flight control method for scheduling a large number of UAV swarm operations according to claim 1, characterized in that: In step S1, the process of finding the optimal path by the energy-saving path planning and obstacle avoidance algorithm for drones is as follows: S11. Initialize the information of the drone cluster. The drone cluster information includes terrain and landform, obstacle positions, climate conditions, and flight missions. The flight missions include initial positions, target routes, and flight times, which are collected through a sensor group. The sensor group includes radar, lidar, and cameras; S12. Generate a pre-planned route. Obtain global information according to the UWB positioning system, and then initially generate a pre-planned path that can connect the starting point and the target point of the drone based on the position of the target point; S13. Generate candidate paths. By sampling the global path, generate a series of candidate paths, and then use the graph search algorithm to search the sampled points to obtain the path with the shortest distance or the smallest cost as the candidate path.

3. A flight control method for scheduling a large number of UAV swarm operations according to claim 2, characterized in that: In step S12, the pre-planned path generation rule is: connect the starting point and the target point; determine whether there is an obstacle intersecting the obtained connection line. If so, select the obstacle closest to the starting point that intersects the connection; select a new path point from the closest obstacle and set it as the starting point.

4. A flight control method for scheduling a large number of UAV swarm operations according to claim 3, characterized in that: In step S13, the process of generating candidate paths is as follows: A1. Create an auxiliary array D to store the shortest path lengths from the starting point to each other vertex; A2. Set the shortest path length of the starting point to 0, and the shortest path lengths of other vertices to infinity; A3. Create a set S to store the visited vertices. Initially, S only contains the starting point; A4. Select the vertex u closest to the starting point from the unvisited vertices, that is, the vertex with the smallest D[u], and add it to the set S; A5. For each neighbor v of vertex u, if the path through u to v is shorter than the currently recorded path from the starting point to v, that is, D[u]+weight(u,v)<D[v], then update D[v] to D[u]+weight(u,v); A6. Repeat the above steps A4 - A5 until all vertices are visited or a certain termination condition is reached.

5. A flight control method for scheduling a large number of UAV swarm operations according to claim 4, characterized in that: In step S2, the cost function is J(i)=∑(w saf f saf (i)+w smo f smo (i)+w ene f ene (i)), where J(i) is the total cost of path i, i is the candidate path index, and f saf is the safety cost function, expressed as the inverse of the minimum distance from the candidate path to the obstacle edge, f smo (i) is the smoothing cost function, represented by the maximum curvature in the candidate path, f ene It is the energy saving cost function, which is expressed by the effective work when the candidate path is executed.

6. A flight control method for scheduling a large number of UAV swarm operations according to claim 5, characterized in that: In step S3, the spatial protection model for cluster drones is a two-layer spatial protection model, which is divided into an alarm area and a warning area. The warning area is a circular area with the centroid of the drone as the center and the minimum safety distance Q as the radius; the warning area is a circular area with the centroid of the drone as the center and 2Q as the radius.

7. A flight control method for scheduling a large number of UAV swarm operations according to claim 6, characterized in that: The spatial protection model for cluster drones receives and processes the flight data of the drone cluster in real time, and judges whether there are potential collision conflicts according to the preset rules or algorithms. When constructing the model, it is necessary to consider the influence of the motion characteristics, flight environment, and communication factors of the drones on conflict detection. The model rules are as follows: S31. When an obstacle appears in the warning area, it is considered as no collision conflict, the detection model is warned, and the dynamic avoidance warning mechanism between drones is activated; S32, when the obstacle appears in the warning area but outside the alarm area, determining the relationship between the speed vector of the obstacle and the speed vector of the drone; S33: When an obstacle appears in the warning area, it is considered a collision. S34: When the distance between the obstacle and the drone is less than one diameter of the drone, it is considered a collision.

8. A flight control method for scheduling a large number of UAV swarm operations according to claim 7, characterized in that: In step S4, the anti-collision strategy includes adjusting the parameters of the flight trajectory, speed or altitude of the drone to avoid potential collision conflicts; the anti-collision algorithm adopts the dynamic frame time slot Aloha algorithm to estimate the tag backlog according to the number of collision time slots by dividing the time slots and dynamically adjusting the frame length.

9. A flight control method for scheduling a large number of UAV swarm operations according to claim 8, characterized in that: The process of estimating the label backlog by the dynamic frame time slot Aloha algorithm is as follows: C1, query time slot, C2, calculate E, S, C at the i-th time slot to get the estimated value n; C3. Determine whether n is within the optimal range. If so, select the next time slot and repeat C2. If not, change the Q value according to the estimated backlog and return to C2.

10. A flight control method for scheduling a large number of UAV swarm operations according to claim 9, characterized in that: In step S5, the UAV collision autonomous flight obstacle avoidance strategy adopts a cluster avoidance strategy with secondary decision-making. The minimum turning radius of the UAV, that is, the maximum turning angle of the UAV, is calculated according to the maximum acceleration of the UAV and the current flight speed. Between the two maximum turning angles on the left and right of the UAV's forward flight direction, with an iteration time interval of 0.1s and an angle interval of 1 degree, the UAV can generate 13 path points at the next moment. A secondary decision is made on the 13 path points generated by the UAV at the next moment to ensure the safe flight of the UAV, wherein the 13 path points include 6 left-turn path points, 6 right-turn path points and 1 straight-line flight path point.

Images