A multi-UAV collaborative path planning method, system, device and storage medium

Abstract

The present invention discloses a multi-UAV collaborative path planning method, system, device and storage medium, which relates to the technical field of UAV path planning, and includes the steps of: obtaining the flight trajectory of multiple UAVs represented by cylindrical coordinates; obtaining a detection function using the increment of the objective function of each path point in the current global optimal path, and obtaining the bottleneck path point of the global optimal path through the probability of each path point in the detection function; obtaining a candidate target vector according to the probability, and adding disturbance to the randomly sampled path, obtaining the candidate path through the current disturbance operator, and generating a new target path point of the global optimal solution; planning the multi-UAV collaborative path through the target path point. The present invention corrects the detected bottleneck waypoints through a two-stage update strategy constructed by operations such as disturbance, and the generated path greatly reduces the existing bottlenecks, generating a high-quality global optimal path with a high success rate.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and in particular to a multi-UAV collaborative path planning method, system, device and storage medium. Background Art

[0002] Unmanned aerial vehicles (UAVs) are autonomous flight systems that do not require a pilot and have attracted widespread attention due to their unique performance and robustness. In the past few decades, UAVs have been widely used in reconnaissance, surveillance, express delivery and rescue missions. Route planning is a key technology in the navigation system of UAVs. Its purpose is to quickly find multiple flight waypoints from the initial position to the destination in the planning space under mission constraints. When faced with complex environments or large areas, multiple UAVs show the advantages of strong robustness and high parallelism when executing complex activity targets. The research on UAVs has received more and more attention from researchers.

[0003] The path planning task is a complex optimization problem under many constraints and has been developed in the past few years. Traditional path planning algorithms mainly include A* algorithm, RRT and artificial potential field (APF) algorithm. However, these traditional methods are difficult to solve path planning problems in complex environments. Compared with traditional algorithms, evolutionary computation algorithms (ECs) can seek feasible and safe paths in complex environments by simulating iterative evolution models of biological populations, such as genetic algorithms (GA), differential evolution algorithms (DE), artificial bee colony algorithms (ABC), particle swarm algorithms (PSO), etc. Due to the advantages of high parallelism, strong robustness and self-organization, EC has shown strong capabilities in planning problems in a more efficient way.

[0004] The particle swarm algorithm (PSO), proposed in 1995, is a common algorithm for ECs and is often used to deal with many optimization problems. Due to its robustness and simplicity, the particle swarm algorithm plays an important role in dealing with path planning problems. However, it also has some disadvantages, such as premature convergence and long optimization time in complex environments, and it is easy to fall into local optimality. When the PSO algorithm is used for UAV route planning, the path obtained has a bottleneck, which seriously affects the quality of the search path. Therefore, path planning in complex environments or large-scale spaces is challenging. Summary of the invention

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and provide a multi-UAV collaborative path planning method, system, device and storage medium to solve the problem that the path obtained during UAV route planning in the prior art has a bottleneck, which seriously affects the quality of the search path.

[0006] The present invention specifically provides the following technical solution: a multi-UAV collaborative path planning method, comprising the following steps:

[0007] Acquire the flight trajectories of multiple UAVs represented by cylindrical coordinates, and obtain a global optimal path through the flight trajectories of multiple UAVs;

[0008] The detection function is obtained by using the increment of the objective function of each path point in the global optimal path, and the bottleneck path point of the global optimal path is obtained by using the probability of each path point in the detection function;

[0009] Add perturbations to the randomly sampled paths, obtain candidate paths through the perturbation operator of the current path, and use the candidate paths to modify the bottleneck path points. In the global optimal path after modifying the bottleneck path points, retain other path points to obtain a new global optimal path.

[0010] The multi-UAV collaborative path is planned through the new global optimal path.

[0011] Preferably, the step of obtaining the flight trajectory of the UAV represented by cylindrical coordinates comprises the following steps:

[0012] Get the Cartesian coordinates of the drone during flight and convert the Cartesian coordinates to cylindrical coordinates q in each route i ; The specific expression is:

[0013] q i =(r i1 ,θ i1 ,z i1 ,r i2 ,θ i2 ,z i2 ,···,r iN ,θ iN ,z iN )

[0014] Among them, (r iN ,θ iN ,z iN ) is the cylindrical coordinate of the Nth path point of the ith particle.

[0015] The velocity corresponding to the cylindrical vector of the cylindrical coordinates is represented by an incremental vector Δq i ; The specific expression is:

[0016] Δq i =(Δr i1 ,Δθ i1 ,Δz i1 ,Δr i2 ,Δθ i2 ,Δz i2 ,···,Δr iN ,Δθ iN ,Δz iN )

[0017] Among them, (Δr iN ,Δθ iN ,Δz iN ) is the velocity representation corresponding to the cylindrical coordinates of the Nth path point of the ith particle;

[0018] The update equation of the planner GTCPSO is:

[0019]

[0020] Among them, Vq ij (t+1) is used to represent the velocity vector of the cylindrical coordinates of the i-th particle, q ij (t+1) is the position vector of the i-th particle in cylindrical coordinates, pb ij (t) is the local optimal position of the ith particle, g j (t) is the position of the global optimal particle in the group, ω is the inertial weight, c1 and c2 are acceleration coefficients, representing the weight of the statistical acceleration term that pushes each particle to the pbest and gbest positions, and r1 and r2 are two random numbers in the range of (0,1).

[0021] Preferably, after obtaining the UAV flight trajectory represented by cylindrical coordinates, the method further includes the following steps:

[0022] The cylindrical vector is converted to Cartesian coordinates to evaluate the quality of the route. The specific expression is:

[0023]

[0024] Preferably, the step of obtaining the detection function by using the increment of the objective function of each path point in the global optimal path comprises the following steps:

[0025] The detection function is obtained through the i-th waypoint of the current global optimal path; the specific expression is:

[0026] ΔF gi =F gi -F g(i-1) (i=2,3···,N)

[0027] Where ΔF gi is the increment of the objective function value of the i-th waypoint on the current global optimal path, F gi and F g(i-1) are the objective function values ​​of the i-th waypoint and the i-1-th waypoint on the current global optimal path respectively.

[0028] Preferably, the method of obtaining the bottleneck path point of the global optimal path by detecting the probability of each path point in the function comprises the following steps:

[0029] Obtain the objective function of the i-th waypoint of the current globally optimal path through the detection function; the specific expression is:

[0030] F gi = a1f gLi + a2f gTi + a3f gHi + a4f gSi + a5f gCi (i = 1, 2, ···, N)

[0031] where f gLi is the route length cost of the i-th waypoint on the current globally optimal path, f gTi is the radar threat cost including the i-th waypoint on the current globally optimal path, f gHi is the altitude cost including the i-th waypoint on the current globally optimal path, f gSi is the smoothness cost including the i-th waypoint on the current globally optimal path, f Ci is the UAV collision cost including the i-th waypoint on the current globally optimal path, and a1 - a5 are the corresponding weight coefficients;

[0032] Determine the bottleneck waypoint of the globally optimal path by obtaining the probability p, and detect the detection function value of each waypoint through the bottleneck waypoint; the specific expression is:

[0033]

[0034] Preferably, add perturbations to the randomly sampled paths, and obtain candidate paths through the perturbation operator of the current path, including the following steps:

[0035] Obtain the current iteration number G and the fixed iteration number G1;

[0036] When G < G1, obtain a homologous targeting vector using the Monte Carlo-based method; when G > G1, obtain candidate paths through the perturbation operator of the current path, and its expression is:

[0037]

[0038] where, r1 is a random number in (0, 1), is the new target vector, Vel Max .r, Vel Max .θ and Vel Max .z are the lower and upper bounds of the velocity, (r i , θ i , z i ) is the current waypoint.

[0039] Preferably, when retaining other path points in the global optimal path after modifying the bottleneck path point to obtain a new global optimal path, the following steps are also included:

[0040] The fitness value of the new path is compared with the fitness value of the original global optimal path before the gene targeting GT operation. If the fitness value of the newly generated path is better than the fitness value of the original global optimal path before the gene targeting GT operation, the newly generated path with a smaller fitness value will replace the original global optimal path as the new global optimal path.

[0041] The present invention provides a multi-UAV collaborative path planning system, comprising:

[0042] A data acquisition module, used for acquiring the flight trajectories of multiple UAVs represented by cylindrical coordinates, and obtaining a global optimal path through the flight trajectories of multiple UAVs;

[0043] A detection module, used to obtain a detection function by using the increment of the objective function of each path point in the global optimal path, and to obtain the bottleneck path point of the global optimal path by using the probability of each path point in the detection function;

[0044] A generation module is used to add perturbations to the randomly sampled paths, obtain candidate paths through the perturbation operator of the current path, and use the candidate paths to modify the bottleneck path points, retain other path points in the global optimal path after modifying the bottleneck path points to obtain a new global optimal path;

[0045] The planning module is used to plan the multi-UAV collaborative path through the new global optimal path.

[0046] The present invention provides a computer device, including a memory and a processor, wherein a program is stored in the memory, and when the program is executed by the processor, the processor executes the steps of the above-mentioned multi-UAV collaborative path planning method.

[0047] The present invention provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned multi-UAV collaborative path planning method are implemented.

[0048] Compared with the prior art, the present invention has the following significant advantages:

[0049] The present invention uses cylindrical coordinates to represent the flight trajectories of multiple UAVs, obtains the flight trajectories of multiple UAVs to obtain the global optimal path, uses the increment of the objective function of each path point in the global optimal path to obtain the detection function, and obtains the bottleneck path point of the global optimal path through the probability of each path point in the detection function, adds disturbances to the randomly sampled path, modifies the bottleneck path point of the obtained candidate path, and retains other path points in the global optimal path after modifying the bottleneck path point to obtain a new global optimal path. The present invention obtains the bottleneck path point of the global optimal path through the detection function obtained by the incremental acquisition of the path point objective function, and corrects the detected bottleneck path point through a two-stage update strategy constructed by operations such as disturbance, so that the generated path greatly reduces the existing bottlenecks and generates a high-quality global optimal path with a high success rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention is an overall flow chart of a multi-UAV collaborative path planning method. DETAILED DESCRIPTION

[0051] The following is a clear and complete description of the technical solutions of the embodiments of the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0052] The purpose of the present invention is to provide a multi-UAV collaborative path planning method, system, device and storage medium. Inspired by gene targeting technology, an improved quadcopter PSO path planning method based on a combination of a cylindrical carrier and gene targeting is specifically proposed. First, a gene detector is used to find the bottleneck path point of the optimal global path; then the bottleneck path point is replaced with the corresponding target vector to improve the quality of the global optimal path. In the process of replacing waypoints, considering the search characteristics of the particle swarm algorithm, a two-stage search strategy is used to generate the corresponding target vector to improve the accuracy of the solution.

[0053] Gene targeting refers to the integration of DNA into a chromosomal target sequence that can accurately replace any gene through homologous recombination mediated by extrachromosomal DNA fragments. GT relies on the crossover of template DNA and genomic target double-stranded DNA. In 2018, Arnold won half of the Nobel Prize in Chemistry for his work on enzyme targeted evolution. This technology is often called gene knockout. The structure of the gene involved is known, but its function is unknown, and experiments are designed to remove the gene at the molecular level or replace it with a similar gene. In 2007, Cap ecchi and Evans won the Nobel Prize for stem cell research, especially the development of gene targeting technology. Weighill et al. performed an interpretive analysis of pancreatic ductal adenocarcinoma (PDAC) and demonstrated the importance of gene targeting scores in identifying complex phenotypic differences. Wang et al. proposed a simple and effective gene-targeted DE (GTDE) algorithm to handle large-scale optimization problems.

[0054] Since the optimization mechanism of the particle swarm algorithm is to use the global optimal solution in all particle learning populations to achieve the update operation, the quality of the global optimal solution will directly affect the efficiency of evolution. Based on this, by introducing gene targeting technology, bottleneck path point detection and targeting vector reconstruction are used to improve the global optimal solution. At the same time, combined with the maneuverability of drones, the cylindrical vector is combined with the particle swarm algorithm.

[0055] The contributions of the present invention mainly include the following two aspects.

[0056] (1) An improved particle swarm planning algorithm based on cylindrical coordinates and gene targeting is proposed for multi-UAV flight path optimization. The method consists of two parts: a detector and a regulator. By identifying the bottleneck path points of the current global optimal path and updating the bottleneck path points, a high-quality global optimal path is generated with a high success rate.

[0057] (2) Considering the particle swarm evolution process, the regulator adopts a two-stage search to solve.

[0058] Problem Description

[0059] Flight space and route representation:

[0060] The flight space is the area where the drone performs flight missions. In three-dimensional space, it is recorded as the waypoint in the flight space, and the polar coordinates of the projection P of the waypoint M are. represents the length of the line segment OM, represents the angle from Ox to OM, is the polar diameter of point M, and is the polar angle of point M, so it is recorded as the cylindrical coordinates of the waypoint M. The waypoint M is expressed in cylindrical coordinates:

[0061] Ω={(r,θ,z)|-r max≤r≤r max ,-θ max ≤θ≤θ max ,-z max ≤z≤z max}

[0062] where represents the scope of the task space, -r max 、r max , -θ max ,θ max 、-z max and z max are the minimum and maximum values ​​of the r, θ, and z coordinates, respectively.

[0063] Objective function:

[0064] The path planning of multiple drones often requires taking off from the starting position at the same time and maintaining a safe distance during the flight to avoid collisions with obstacles and threat areas, and finally reach the target location. In the trajectory planning of multiple drones, factors such as trajectory length, terrain collision, and radar threat need to be considered. In addition, the spatial collision between multiple drones needs to be considered, that is, maintaining an appropriate safe distance between drones during flight. Therefore, the objective function is:

[0065] f=a1f L +a2f T +a3f H +a4f S +a5f C

[0066] Where f L is the route length cost, f T is the radar threat cost, f H is the height cost, f S is the smoothing cost, f C is the drone collision cost, and a1-a5 are weight coefficients.

[0067] Since the path obtained by the algorithm has bottlenecks, the quality of the search path is seriously affected, so path planning in complex environments or large-scale spaces is challenging. Since the optimization mechanism of the particle swarm algorithm is to use the global optimal solution in all particle learning populations to achieve the update operation, the quality of the global optimal solution will directly affect the efficiency of evolution. Therefore, by introducing gene targeting technology, the global optimal solution is improved through bottleneck path point detection and targeting vector reconstruction. At the same time, the mobility of drones is used to combine cylindrical vectors into the particle swarm algorithm.

[0068] Based on this, the multi-UAV collaborative path planning method proposed in the present invention solves the problems of the prior art and specifically includes the following steps:

[0069] Step S1: Obtain the flight trajectories of multiple UAVs represented by cylindrical coordinates, and obtain the global optimal path through the flight trajectories of multiple UAVs.

[0070] Initialize the parameters of GTCPSO, generate the initial paths represented by the cylindrical coordinates of N drones, convert each initial path into Cartesian coordinates, calculate the fitness value of each path, and find the optimal path through the fitness value.

[0071] Specifically:

[0072] Obtaining the flight trajectory of a drone expressed in cylindrical coordinates includes the following steps:

[0073] Get the Cartesian coordinates of the drone during flight and convert the Cartesian coordinates into cylindrical coordinates in each route; the specific expression is:

[0074] q i =(r i1 ,θ i1 ,z i1 ,r i2 ,θ i2 ,z i2 ,···,r iN ,θ iN ,z iN )

[0075] Among them, (r iN ,θ iN ,z iN ) is the cylindrical coordinate of the Nth path point of the ith particle. That is, the use of cylindrical coordinates to represent the flight trajectory of the drone can better meet the dynamic characteristics of the drone. In three-dimensional space, each route is represented by cylindrical coordinates instead of the original Cartesian coordinates. Assume that the cylindrical coordinates of the waypoint are (r iN ,θ iN ,z iN ), then the flight route consisting of N waypoints can be represented by a cylindrical vector in three-dimensional space.

[0076] The velocity corresponding to the cylindrical vector of the cylindrical coordinates is expressed by an incremental vector; the specific expression is:

[0077] Δq i =(Δr i1 ,Δθ i1 ,Δz i1 ,Δr i2 ,Δθ i2 ,Δz i2 ,···,Δr iN ,Δθ iN ,Δz iN )

[0078] Among them, (ΔriN ,Δθ iN ,Δz iN ) is the velocity representation corresponding to the cylindrical coordinates of the Nth path point of the ith particle.

[0079] The update equation of the planner GTCPSO is:

[0080]

[0081] Among them, Vq ij (t+1) is used to represent the velocity vector of the cylindrical coordinates of the i-th particle, q ij (t+1) is the position vector of the cylindrical coordinates of the ith particle, pb ij (t) is the local optimal position of the ith particle, g j (t) is the position of the global optimal particle in the group, ω is the inertial weight, c1 and c2 are acceleration coefficients, representing the weight of the statistical acceleration term that pushes each particle to the pbest and gbest positions, and r1 and r2 are two random numbers in the range of (0,1).

[0082] After obtaining the UAV flight trajectory represented by cylindrical coordinates, the following steps are also included:

[0083] Convert the cylindrical vector to Cartesian coordinates to evaluate the quality of the route; the specific expression is:

[0084]

[0085] Step S2: Utilize the increment of the objective function of each waypoint in the current global optimal path to obtain the detection function, and obtain the bottleneck waypoint of the global optimal path through the probability of each waypoint in the detection function.

[0086] Based on the GT idea, a cylindrical coordinate particle swarm algorithm based on GT technology is proposed, which further improves the solution accuracy of the particle swarm algorithm in multi-UAV control path planning. In the particle swarm algorithm, the position update of all particles depends on the guidance of the global optimal solution, so the quality of the global optimal solution affects the solution accuracy of the particle swarm algorithm. The introduction of GT technology to optimize the global optimal path effectively improves the quality of the global optimal path.

[0087] When using GT technology to deal with path planning problems, first of all, the bottleneck path point will directly affect the current global optimal path and can be regarded as the pathogenic gene; the bottleneck path point of the constraint algorithm to find the global optimal solution can be regarded as the pathogenic gene. According to the objective functions of multiple drones, a simple and effective detection function is established to detect the bottleneck waypoints of the global optimal path.

[0088] Specifically:

[0089] The detection function is obtained by using the increment of the objective function of each waypoint in the current global optimal path, including the following steps:

[0090] The detection function is obtained through the i-th waypoint of the current global optimal path; the specific expression is:

[0091] ΔF gi =F gi -F g(i-1) (i=2,3···,N)

[0092] Where ΔF gi is the increment of the objective function value of the i-th waypoint on the current global optimal path, F gi and F g(i-1) They are the objective function values ​​of the i-th waypoint and the (i-1)-th waypoint on the current global optimal path respectively.

[0093] Obtaining the bottleneck waypoint of the global optimal path by detecting the probability of each waypoint in the function also includes the following steps:

[0094] The objective function of the current global optimal path i is obtained through the detection function; the specific expression is:

[0095] F gi =a1f gLi +a2f gTi +a3f gHi +a4f gSi +a5f gCi (i=1,2,···,N)

[0096] where f gLi is the path length cost of the i-th waypoint on the current global optimal path, f gTi is the radar threat cost of the i-th waypoint on the current global optimal path, f gHi is the height cost of the i-th waypoint on the current global optimal path, f gSi is the smoothing cost of the i-th waypoint on the current global optimal path, f Ci is the collision cost of the UAV containing the i-th waypoint on the current global optimal path, and a1-a5 are the corresponding weight coefficients;

[0097] The bottleneck waypoint of the global optimal path is determined by obtaining the probability p, and the detection function value of each waypoint is detected through the bottleneck waypoint; the specific expression is:

[0098]

[0099] Step S3: Add perturbations to the randomly sampled paths, obtain candidate paths through the perturbation operator of the current path, and use the candidate paths to modify the bottleneck path points. Retain other path points in the globally optimal path after modifying the bottleneck path points to obtain a new globally optimal path.

[0100] Specifically:

[0101] Use DNA recombination technology to construct new targeted waypoints. Here, based on GT technology, only the bottleneck waypoints of the globally optimal path are reconstructed, and this method will be simpler and more efficient. To increase the diversity of the algorithm, the Monte Carlo method is adopted to select some particles from the current particles as candidate particles according to the given sample probability. At the same time, it is considered that the particle swarm algorithm is likely to approach the global optimal solution during the evolution process. For this reason, a two-stage strategy is designed to generate the target vector through the two-stage strategy, put the target vector into the optimal path to replace the bottleneck path points, select the new globally optimal path, and update the globally optimal path. When the end condition is met, the optimal paths of multiple UAVs are output, otherwise recalculate to increase the diversity of the global optimal solution.

[0102] First, adopt the Monte Carlo method to obtain candidate target vectors using probability; then considering that the particle swarm optimization algorithm may cause a large number of particles to converge to the global optimal solution at a certain stage of the evolution process, add perturbations to the paths randomly sampled in the group to help the group jump out of the local optimal solution.

[0103] Let G represent the current iteration number, and G1 be the fixed iteration number. When G < G1, this paper adopts the Monte Carlo-based method to obtain a homologous targeted vector with a probability of. When the current generation G > G1, candidate paths are obtained through the perturbation operator of the current path, and its expression is:

[0104]

[0105] After constructing the target vector, m1 new candidate routes are obtained. Suppose the number of these candidate paths is m1. The bottleneck waypoints of the globally optimal path P g will be modified by the waypoints corresponding to these candidate paths, while other waypoints in the globally optimal path will be retained. The mutated embryonic stem cells, that is, the newly generated paths, are recorded as P' i . The above three steps constitute a complete search process based on GT technology. Compare the fitness value of the new path with the fitness value of the original globally optimal path before the GT operation. If the fitness value of the newly generated path is better than the fitness value of the original globally optimal path before the GT operation, the newly generated path with the smaller fitness value will replace the original globally optimal path as the new globally optimal path.

[0106] Step S4: Plan the cooperative paths of multiple UAVs through the new globally optimal path.

[0107] Specifically:

[0108] The bottleneck waypoints in the global optimal path are detected using the waypoint detection function, and a new target vector is constructed through a two-stage strategy, which provides more promising examples for the evolution of the algorithm and enhances the potential for the development of the flight space.

[0109] Based on the above method, the present invention provides a multi-UAV collaborative path planning system, including: a data acquisition module, a detection module, a generation module and a planning module.

[0110] Among them, the data acquisition module is used to obtain the flight trajectories of multiple UAVs represented by cylindrical coordinates, and obtain the global optimal path through the flight trajectories of multiple UAVs; the detection module is used to obtain the detection function using the increment of the objective function of each path point in the global optimal path, and obtain the bottleneck path point of the global optimal path through the probability of each path point in the detection function; the generation module is used to add perturbations to the randomly sampled paths, obtain candidate paths through the perturbation operator of the current path, and use the candidate paths to modify the bottleneck path points, and retain other path points in the global optimal path after modifying the bottleneck path points to obtain a new global optimal path; the planning module is used to plan the collaborative path of multiple UAVs through the new global optimal path.

[0111] The present invention also provides a computer device, including a memory and a processor, wherein a program is stored in the memory, and when the program is executed by the processor, the processor executes the steps of a multi-UAV collaborative path planning method.

[0112] According to the disclosed embodiments, a computing device may communicate with one or more external devices (e.g., a keyboard, a pointing device, Bluetooth communications, etc.), or with any device (e.g., a router, a modem, etc.) that enables a computing device to communicate with one or more other computing devices.

[0113] The present invention also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, a method for predicting the remaining time of an aircraft is implemented.

[0114] According to the disclosed embodiments, the storage medium may be a non-volatile computer-readable storage medium, such as but not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0115] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0116] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A multi-UAV collaborative path planning method, characterized in that: Including: Obtain the multi-UAV flight trajectory represented by cylindrical coordinates, and obtain the globally optimal path through the multi-UAV flight trajectory; Obtain the detection function by using the increment of the objective function of each path point in the globally optimal path, and obtain the bottleneck path points of the globally optimal path through the probability of each path point in the detection function; Add perturbations to the randomly sampled path, obtain the candidate path through the perturbation operator of the current path, and use the candidate path to modify the bottleneck path points, and retain other path points in the globally optimal path after modifying the bottleneck path points to obtain a new globally optimal path; Plan the multi-UAV cooperative path through the new globally optimal path; The obtaining the detection function by using the increment of the objective function of each path point in the globally optimal path includes the following steps: Obtain the detection function through the u-th path point of the current globally optimal path; the specific expression is: ΔF gu =F gu -F g(u-1) Where, ΔF gu is the increment of the objective function value of the u-th waypoint on the current global optimal path, u=2,3···,N,F gu and F g(u-1) are the objective function values ​​of the u-th waypoint and the u-1-th waypoint of the current global optimal path respectively; The obtaining the bottleneck path points of the globally optimal path through the probability of each path point in the detection function includes the following steps: Obtain the objective function of the h-th waypoint on the current globally optimal path through the detection function; the specific expression is: <h2 style=";text-align:left;direction:ltr">F<h2 style=";text-align:left;direction:ltr"> gh <h2 style=";text-align:left;direction:ltr"> =a1f<h2 style=";text-align:left;direction:ltr"> gLh <h2 style=";text-align:left;direction:ltr"> +a2f<h2 style=";text-align:left;direction:ltr"> gTh <h2 style=";text-align:left;direction:ltr"> +a3f<h2 style=";text-align:left;direction:ltr"> gHh <h2 style=";text-align:left;direction:ltr"> +a4f<h2 style=";text-align:left;direction:ltr"> gSh <h2 style=";text-align:left;direction:ltr"> +a5f<h2 style=";text-align:left;direction:ltr"> gCh Among them, f gLh is the path length cost of the h-th waypoint on the current global optimal path, f gTh is the radar threat cost of the h-th waypoint on the current global optimal path, f gHh is the height cost of the h-th waypoint on the current global optimal path, f gSh is the smoothing cost of the h-th waypoint on the current global optimal path, f Ch is the collision cost of the UAV containing the h-th waypoint on the current global optimal path, a1-a5 are the corresponding weight coefficients; h=1,2,···,N; Determine the bottleneck path points of the globally optimal path by obtaining the probability p, and detect the detection function values of each path point through the bottleneck path points; the specific expression is: The adding perturbations to the randomly sampled path and obtaining the candidate path through the perturbation operator of the current path includes the following steps: Obtain the current iteration number G and the fixed iteration number G1; When G < G1, obtain a homologous targeting vector by using the Monte Carlo-based method; when the current generation G > G1, obtain the candidate path through the perturbation operator of the current path, and its expression is: Among them, r1 is a random number (0,1), is the new target vector, Vel Max .r、Vel Max .θ and Vel Max .z is the lower and upper bounds of the velocity, (r i ,θ i ,z i ) is the current waypoint.

2. A multi-UAV collaborative path planning method as claimed in claim 1, characterized in that: The obtaining the UAV flight trajectory represented by cylindrical coordinates includes the following steps: Get the Cartesian coordinates of the drone during flight and convert the Cartesian coordinates to cylindrical coordinates q in each route i ; The specific expression is: q i =(r i1 ,θ i1 ,z i1 ,r i2 ,θ i2 ,z i2 ,···,r iN ,θ iN ,z iN ) Among them, (r iN ,θ iN ,z iN ) is the cylindrical coordinate of the Nth path point of the ith particle; The velocity corresponding to the cylindrical vector of the cylindrical coordinates is represented by an incremental vector Δq i ; The specific expression is: Δq i =(Δr i1 ,Δθ i1 ,Δz i1 ,Δr i2 ,Δθ i2 ,Δz i2 ,···,Δr iN ,Δθ iN ,Δz iN ) Among them, (Δr iN ,Δθ iN ,Δz iN ) is the velocity representation corresponding to the cylindrical coordinates of the Nth path point of the ith particle; The update equation of the planner GTCPSO is: Where i is the number of particles, i = 1, 2, ..., m, j = 1, 2, ... N, Vq ij (t+1) is used to represent the velocity vector of the cylindrical coordinates of the i-th particle, q ij (t+1) is the position vector of the i-th particle in cylindrical coordinates, pb ij (t) is the local optimal position of the ith particle, g j (t) is the position of the global optimal particle in the group, ω is the inertial weight, c1 and c2 are acceleration coefficients, representing the weight of the statistical acceleration term that pushes each particle to the pbest and gbest positions, and r1 and r2 are two random numbers in the range of (0,1).

3. A multi-UAV collaborative path planning method as claimed in claim 2, characterized in that: After obtaining the UAV flight trajectory represented by cylindrical coordinates, the following steps are further included: Convert the cylindrical vector to Cartesian coordinates and evaluate the quality of the route; the specific expression is:

4. The method for multi-UAV collaborative path planning according to claim 1, characterized in that: When retaining other path points in the globally optimal path after modifying the bottleneck path points to obtain a new globally optimal path, the following steps are further included: Compare the fitness value of the new path with the fitness value of the original globally optimal path before the gene targeting GT operation. If the fitness value of the newly generated path is better than the fitness value of the original globally optimal path before the gene targeting GT operation, the newly generated path with the smaller fitness value will replace the original globally optimal path as the new globally optimal path.

5. A multi-UAV collaborative path planning system, characterized in that: Including: A data acquisition module, configured to obtain the multi-UAV flight trajectory represented by cylindrical coordinates, and obtain the globally optimal path through the multi-UAV flight trajectory; A detection module, configured to obtain the detection function by using the increment of the objective function of each path point in the globally optimal path, and obtain the bottleneck path points of the globally optimal path through the probability of each path point in the detection function; A generation module, configured to add perturbations to the randomly sampled path, obtain the candidate path through the perturbation operator of the current path, and use the candidate path to modify the bottleneck path points, and retain other path points in the globally optimal path after modifying the bottleneck path points to obtain a new globally optimal path; A planning module for planning the cooperative paths of multiple unmanned aerial vehicles through the new globally optimal path; The obtaining of the detection function by using the increment of the objective function of each path point in the globally optimal path includes the following steps: Obtaining the detection function through the u-th path point of the current globally optimal path; the specific expression is: ΔF gu =F gu -F g(u-1) Where, ΔF gu is the increment of the objective function value of the u-th waypoint on the current global optimal path, u=2,3···,N,F gu and F g(u-1) are the objective function values ​​of the u-th waypoint and the u-1-th waypoint of the current global optimal path respectively; The obtaining of the bottleneck path points of the globally optimal path by using the probability of each path point in the detection function includes the following steps: Obtaining the objective function of the h-th waypoint on the current globally optimal path through the detection function; the specific expression is: <h2 style=";text-align:left;direction:ltr">F<h2 style=";text-align:left;direction:ltr"> gh <h2 style=";text-align:left;direction:ltr"> =a1f<h2 style=";text-align:left;direction:ltr"> gLh <h2 style=";text-align:left;direction:ltr"> +a2f<h2 style=";text-align:left;direction:ltr"> gTh <h2 style=";text-align:left;direction:ltr"> +a3f<h2 style=";text-align:left;direction:ltr"> gHh <h2 style=";text-align:left;direction:ltr"> +a4f<h2 style=";text-align:left;direction:ltr"> gSh <h2 style=";text-align:left;direction:ltr"> +a5f<h2 style=";text-align:left;direction:ltr"> gCh Among them, f gLh is the path length cost of the h-th waypoint on the current global optimal path, f gTh is the radar threat cost of the h-th waypoint on the current global optimal path, f gHh is the height cost of the h-th waypoint on the current global optimal path, f gSh is the smoothing cost of the h-th waypoint on the current global optimal path, f Ch is the collision cost of the UAV containing the h-th waypoint on the current global optimal path, a1-a5 are the corresponding weight coefficients; h=1,2,···,N; Determining the bottleneck path points of the globally optimal path by obtaining the probability p, and detecting the detection function values of each path point through the bottleneck path points; the specific expression is: Adding perturbations to the randomly sampled paths and obtaining candidate paths through the perturbation operator of the current path includes the following steps: Obtaining the current iteration number G and the fixed iteration number G1; When G < G1, obtaining a homologous targeting vector by using the Monte Carlo-based method; when the current generation G > G1, obtaining candidate paths through the perturbation operator of the current path, and its expression is: Among them, r1 is a random number (0,1), is the new target vector, Vel Max .r、Vel Max .θ and Vel Max .z is the lower and upper bounds of the velocity, (r i ,θ i ,z i ) is the current waypoint.

6. A computer device, characterized in that: Including a memory and a processor, wherein a program is stored in the memory, and when the program is executed by the processor, the processor executes the steps of a method for planning cooperative paths of multiple unmanned aerial vehicles according to any one of claims 1 to 4.

7. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of a method for planning cooperative paths of multiple unmanned aerial vehicles according to any one of claims 1 to 4.

Images