A fast planning method for multi-UAV regional coverage paths

Abstract

The present invention relates to a method for fast planning of paths for covering a region by multiple unmanned aerial vehicles, and relates to the field of control or regulation systems for non-electrical variables. The method comprises: using an intelligent path planning model to intelligently predict the flight duration corresponding to a certain flight path based on the three-dimensional coordinate data, various line data, and multiple flight configuration information corresponding to each spatial point of the flight region covered by a certain flight path used by the current unmanned aerial vehicle group to complete the target flight route; and using the flight path corresponding to the minimum value of each flight duration as the planned flight path. Through the present invention, in view of the technical problem that it is difficult to take into account both the minimization of the total flight time and the obstacle constraints, and the high cost of path planning calculation and time costs, an intelligent path planning model is used to traverse the flight durations corresponding to each flight path simulated by the current unmanned aerial vehicle group to complete the target flight route based on multiple basic data, thereby solving the above technical problems.

Description

Technical Field

[0001] The present invention relates to the field of control or regulation systems for non-electrical variables, and in particular to a method for rapid planning of regional coverage paths for multiple unmanned aerial vehicles. Background Art

[0002] A drone swarm refers to a group of multiple drones that work together. It can be used in various application fields such as aerial photography, performances, mapping, rescue, plant protection, and airspace lifting. The control of a drone swarm involves communication and coordination between multiple drones, and a control algorithm is required to manage the flight of drones. A common control method is the centralized control method, in which a master computer or controller is responsible for managing the entire drone swarm. It collects status information of all drones and determines the best action for each drone. Another control method is the decentralized control method, in which each drone has its own controller and does not need to communicate with other drones. Each drone determines its own action based on its own status and sensor information. No matter how it is controlled, the path planning of a drone swarm, i.e., multiple drones, when performing a task determines the degree and quality of task completion, and is therefore crucial.

[0003] For example, the Chinese invention patent with the patent publication number CN116225044A proposes a UAV coverage path planning algorithm for complex environment areas. First, the complex environment area is discretized into several square grids according to the size of obstacles in the complex environment area and the shooting range of the onboard camera sensor. Then, a novel Gadget-based graph model is established. Among them, a Gadget contains eight nodes, which represent the eight movement directions of the UAV in the horizontal plane. An undirected edge is established between any two nodes in each Gadget. The undirected edge and the edge weight respectively represent the turning process of the UAV and the cost of turning. A directed edge is established between two nodes in adjacent Gadgets. The directed edge and the edge weight respectively represent the straight flight stage of the UAV and the cost of straight flight. Finally, the problem is transformed into an instance of the generalized traveling salesman problem (GTSP) and solved using the GLNS solver.

[0004] For example, the Chinese invention patent with patent publication number CN118778692A proposes a method for fast planning of multi-UAV regional coverage paths based on smart airports. This method is aimed at scenarios where multiple UAVs with single flight endurance constraints and using smart airports for charging or battery replacement for regional coverage use integer linear programming modeling to minimize the total flight time under the constraints of UAV endurance and full regional coverage. At the same time, the present invention also discloses a fast optimization algorithm for the above integer linear programming, combining aggregation and block and random improvement methods.

[0005] It can be seen that the above two solutions in the prior art, when providing solutions for the regional coverage path planning of multiple UAVs, either only consider the constraints of obstacles without considering the minimization of the total flight time, or only consider the minimization of the total flight time without considering the constraints of obstacles. There is a lack of technical solutions that take into account both the constraints of obstacles and the minimization of the total flight time. Moreover, the above two solutions in the prior art need to rely on complex algorithms or complex models to provide solutions for the regional coverage path planning of multiple UAVs. The establishment process of these complex algorithms or complex models is relatively slow and cumbersome. At the same time, using these complex algorithms or complex models to find the optimization and planning of the regional coverage path of multiple UAVs is time-consuming and labor-intensive, and requires a lot of computing costs and time costs. Summary of the invention

[0006] In order to solve the technical problems in the prior art, the present invention provides a method for rapid planning of paths for regional coverage by multiple UAVs. Before the current UAV group performs actual flight, an intelligent path planning model with a customized structure is used to traverse the flight times corresponding to the flight paths simulated by the current UAV group to complete the target flight route based on multiple basic data that are targeted and screened. The flight path corresponding to the minimum value of the flight times is used as the planned flight path for the current UAV group to complete the target flight route and is sent to the remote control device of the current UAV group to configure the actual flight of the current UAV group, thereby completing the rapid and reliable planning of the optimal flight path of the UAV group, while effectively bypassing various obstacle points on the target flight route, and shortening the flight time of the UAV group as much as possible. More importantly, the above-mentioned intelligent mechanism reduces the computational cost and time cost of path planning, and has a relatively broad application space.

[0007] According to the present invention, a method for rapid planning of multi-UAV area coverage paths is provided, the method comprising:

[0008] Acquire various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route;

[0009] Obtain the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route, so as to output multiple flight configuration information of the current drone group;

[0010] Performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after the multiple training actions and outputting it as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm;

[0011] Input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and run the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model;

[0012] The flight path corresponding to the minimum value of each time duration corresponding to each flight path used by the current drone group to complete the target flight route is sent to the remote control device of the current drone group as the planned flight path for the current drone group to complete the target flight route.

[0013] It can be seen that the present invention has at least the following four key inventive concepts:

[0014] Invention concept 1: Before the current drone group performs actual flight, an intelligent path planning model is used to traverse the flight durations corresponding to the various flight paths simulated by the current drone group to complete the target flight route, and the flight path corresponding to the minimum value of the various flight durations is sent to the remote control device of the current drone group as the planned flight path for the current drone group to complete the target flight route to configure the actual flight of the current drone group, thereby completing the rapid and reliable planning of the optimal flight path of the drone group, while effectively bypassing various obstacle points on the target flight route, and shortening the flight time of the drone group as much as possible;

[0015] Inventive concept 2: A customized intelligent path planning model is used to intelligently predict the flight time corresponding to each flight path simulated by the current drone group to complete the target flight route. The structural customization of the intelligent path planning model is manifested in that the intelligent path planning model is a convolutional neural network after multiple training actions, and the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone group, thereby ensuring the stability and effectiveness of the prediction of the flight time of each flight path to avoid each obstacle point on the target flight route;

[0016] Inventive concept three: providing multiple sets of targeted and screened basic data for intelligent prediction of flight duration, the multiple sets of basic data include three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various line data of the target flight route that the current drone group is about to complete, and multiple flight configuration information of the current drone group. Specifically, the various line data of the target flight route that the current drone group is about to complete are the three-dimensional coordinate data of the flight starting point set of the target flight route, the three-dimensional coordinate data of the flight end point set of the target flight route, and multiple sets of three-dimensional coordinate data corresponding to multiple obstacle points between the flight starting point set and the flight end point set of the target flight route. The multiple flight configuration information of the current drone group is the set flight speed when the current drone group executes the target flight route, the radius of the minimum sphere that can accommodate the flight starting point set and the flight end point set of the target flight route in space, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route, thereby further ensuring the stability and effectiveness of the prediction of the flight duration of each flight path to avoid each obstacle point on the target flight route;

[0017] Inventive concept four: In each training action performed on the convolutional neural network, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete a past flight route, the route data of the past flight route completed by the current drone group, the set flight speed when the current drone group executes the past flight route, the radius of the smallest sphere that can accommodate the flight starting point set and the flight end point set of the past flight route in space, the number of drones in the current drone group, and the longest distance from the flight starting point set of the past flight route to the flight end point set of the past flight route are used as various input data of the convolutional neural network, and the actual time used by the current drone group to complete the past flight route using the certain flight path is used as the single output data of the convolutional neural network to complete this training action of the convolutional neural network, thereby ensuring the training effect of each training action of the convolutional neural network. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

[0019] Figure 1 The present invention is a technical flow chart of a method for rapid planning of multi-UAV area coverage paths.

[0020] Figure 2 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to a first embodiment of the present invention.

[0021] Figure 3 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to a second embodiment of the present invention.

[0022] Figure 4 The present invention is a flowchart showing a method for rapid planning of paths for multiple UAV area coverage according to a third embodiment of the present invention.

[0023] Figure 5 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to the fourth embodiment of the present invention.

[0024] Figure 6 The present invention is a flowchart showing a method for quickly planning paths for multiple UAV area coverage according to the fifth embodiment of the present invention. DETAILED DESCRIPTION

[0025] like Figure 1 As shown, a technical flow chart of a method for rapid planning of multi-UAV area coverage paths according to the present invention is given.

[0026] like Figure 1 As shown, the specific technical process of the present invention is as follows:

[0027] Technical process A: Assign the current drone group a flight mission to fly the target flight route, and simulate various flight paths for the target flight route before the actual flight, providing path data for the subsequent planning of the optimal flight path that takes the shortest time and effectively avoids multiple obstacles;

[0028] For example, the target flight route is a straight line trajectory from each flight starting point to each flight end point, and the number of flight starting points is the same as the number of flight end points and is equal to the number of multiple drones in the current drone group;

[0029] The difference is that although each flight path also starts from each flight starting point of the target flight route to each flight end point, it is not necessarily a straight line trajectory, and it is necessary to avoid each obstacle point between each flight starting point and each flight end point of the target flight route;

[0030] Specifically, Figure 1 As shown, the judgment criteria for each obstacle point between each flight starting point and each flight end point of the target flight route is: multiple obstacle points existing in the smallest sphere that can accommodate each flight starting point and each flight end point of the target flight route in space;

[0031] The key is that Figure 1 The four obstacle points in the figure are illustrative. In fact, the number of obstacle points on the target flight path is not fixed and the distribution pattern is random. No matter how many obstacle points are on the target flight path and how these obstacle points are distributed, they will not affect the subsequent prediction results and analysis results of the flight duration of each flight path of the present invention.

[0032] Technical process B: Build a customized intelligent path planning model for the intelligent prediction of the flight time of each flight path, thereby ensuring the stability and effectiveness of the prediction of the flight time of each flight path to avoid obstacles on the target flight route;

[0033] Specifically, the structural customization of the intelligent path planning model is as follows: the intelligent path planning model is a convolutional neural network after multiple training actions, and the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone group;

[0034] More specifically, in each training action performed on the convolutional neural network, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete a certain past flight route, the route data of the past flight route that the current drone group has completed, the set flight speed when the current drone group executes the past flight route, the radius of the minimum sphere that can accommodate the flight starting point set and the flight end point set of the past flight route in space, the number of drones in the current drone group, and the longest distance from the flight starting point set of the past flight route to the flight end point set of the past flight route are used as various input data of the convolutional neural network, and the actual time used by the current drone group to complete the past flight route using the certain flight path is used as the single output data of the convolutional neural network to complete this training action of the convolutional neural network, thereby ensuring the training effect of each training action of the convolutional neural network;

[0035] Technical process C: Providing targeted and screened multiple basic data for the intelligent prediction of the flight duration of each flight path, thereby further ensuring the stability and effectiveness of the prediction of the flight duration of each flight path to avoid obstacles on the target flight route;

[0036] Specifically, the multiple basic data include three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group;

[0037] More specifically, the various route data of the target flight route that the current drone group is about to complete are the three-dimensional coordinate data of the flight starting point set of the target flight route, the three-dimensional coordinate data of the flight end point set of the target flight route, and multiple sets of three-dimensional coordinate data corresponding to multiple obstacle points between the flight starting point set and the flight end point set of the target flight route, and the multiple flight configuration information of the current drone group are the set flight speed when the current drone group executes the target flight route, the radius of the minimum sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route;

[0038] Technical process D: The intelligent path planning model customized by technical process B is used to intelligently predict the flight duration of each flight path simulated in technical process A based on multiple basic data selected by technical process C;

[0039] Technical process E: Based on technical process D, the flight durations corresponding to the flight paths of the target flight route are obtained, and the flight path corresponding to the minimum flight duration is used as the optimal flight path that takes into account the lowest flight time cost and effectively avoids various obstacles on the target flight route;

[0040] In this way, by selecting the optimal flight path, the various time-sharing positions of multiple drones that make up the drone swarm during the actual flight process can be quickly planned, and the various time-sharing positions constitute the coverage area of ​​multiple drones, thereby completing the rapid planning of the multi-drone area coverage path, while effectively bypassing various obstacle points on the target flight route, shortening the flight time of the drone swarm as much as possible.

[0041] The key points of the present invention are: the definition of the smallest sphere that can spatially accommodate each flight starting point and each flight end point of the target flight route, the judgment of multiple obstacle points on the target flight route, the customized structure of the intelligent path planning model, the design of a single training mechanism, and the targeted screening of multiple basic data.

[0042] Below, a method for rapid planning of multi-UAV area coverage paths of the present invention will be specifically described in the form of an embodiment.

[0043] First embodiment

[0044] Figure 2 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to a first embodiment of the present invention.

[0045] like Figure 2 As shown, the method for rapid planning of multi-UAV area coverage paths includes the following steps:

[0046] Step S21: obtaining various route data of a target flight route that is about to be completed by the current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route;

[0047] For example, the target flight route is a straight line trajectory from each flight starting point to each flight end point, each flight starting point can be numerically represented by each three-dimensional coordinate data corresponding to each flight starting point, each flight end point can be numerically represented by each three-dimensional coordinate data corresponding to each flight end point, and the number of flight starting points is the same as the number of flight end points and is equal to the number of multiple drones in the current drone group;

[0048] It can be seen that the numerical representation of the target flight route can actually be regarded as the three-dimensional spatial numerical value of each straight line from each three-dimensional coordinate data corresponding to each flight starting point to each three-dimensional coordinate data corresponding to each flight end point;

[0049] Step S22: obtaining the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route, so as to output multiple flight configuration information of the current drone group;

[0050] For example, obtaining the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route as multiple flight configuration information outputs of the current drone group includes: multiple data acquisition components can be used to respectively obtain the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route;

[0051] Step S23: performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after multiple training actions and outputting it as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm;

[0052] In fact, by performing multiple training actions on the convolutional neural network to complete multiple redefinitions of the internal structure of the multiple training actions, a data foundation is laid for the reliability and stability of the flight duration prediction performed by the convolutional neural network obtained after multiple training actions, i.e., the intelligent path planning model;

[0053] Specifically, the number of training actions of the convolutional neural network and the number of drones in the current drone swarm are monotonically positively correlated, including: the number of drones in the current drone swarm is 10, the number of training actions of the convolutional neural network is 150, the number of drones in the current drone swarm is 15, the number of training actions of the convolutional neural network is 200, the number of drones in the current drone swarm is 20, the number of training actions of the convolutional neural network is 250, the number of drones in the current drone swarm is 25, the number of training actions of the convolutional neural network is 300, and so on;

[0054] Step S24: inputting the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and running the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model;

[0055] The difference is that although each flight path also starts from each flight starting point of the target flight route to each flight end point, it is not necessarily a straight line trajectory, and it is necessary to avoid each obstacle point between each flight starting point and each flight end point of the target flight route;

[0056] Specifically, the judgment criteria for each obstacle point between each flight starting point and each flight end point of the target flight route are: multiple obstacle points existing in the smallest sphere that can spatially accommodate each flight starting point and each flight end point of the target flight route;

[0057] Step S25: sending the flight path corresponding to the minimum value of each duration corresponding to each flight path used by the current drone group to complete the target flight route as the planned flight path for the current drone group to complete the target flight route to the remote controller of the current drone group;

[0058] In this way, before the current drone group performs the actual flight mission of the target flight route, the flight duration of each flight trajectory of the target flight route that avoids multiple obstacles can be effectively predicted through the artificial intelligence mechanism, and the shortest flight duration is allocated to the current drone group as the optimal flight duration, thereby realizing the rapid planning of the flight path that takes into account multiple obstacles at random locations and the flight time cost;

[0059] In this way, by selecting the optimal flight path, the time-sharing positions of multiple drones constituting a drone group during the actual flight process are quickly planned, and the time-sharing positions constitute the coverage area of ​​multiple drones, thereby actually completing the rapid planning of the coverage path of multiple drones.

[0060] Among them, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route to be completed by the current drone group, and the multiple flight configuration information of the current drone group are input into the intelligent path planning model, and the intelligent path planning model is run to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model, including: each spatial point of the flight area covered by any flight path used by the current drone group to complete the target flight route does not include multiple obstacle points between the flight starting point set and the flight end point set of the target flight route, so as to avoid the multiple obstacle points;

[0061] Specifically, each spatial point of the flight area covered by any flight path used by the current drone group to complete the target flight route does not include multiple obstacle points between the flight starting point set and the flight end point set of the target flight route, so as to avoid the multiple obstacle points including: there is no regularity in the distribution of the positions of the obstacle points on different flight routes;

[0062] Wherein, obtaining various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data respectively corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route, including: the plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route are a plurality of obstacle points existing in a minimum sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route;

[0063] Among them, performing multiple training actions on the convolutional neural network to obtain the convolutional neural network after multiple training actions and outputting it as an intelligent path planning model, the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone group, including: in each training action performed on the convolutional neural network, using the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete a certain past flight route, the various route data of the past flight route completed by the current drone group, the set flight speed when the current drone group executes the past flight route, the radius of the minimum sphere that can spatially accommodate the flight starting point set and the flight end point set of the past flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the past flight route to the flight end point set of the past flight route as the various input data of the convolutional neural network, using the actual time used by the current drone group to complete the past flight route using the certain flight path as the single output data of the convolutional neural network, and completing this training action of the convolutional neural network;

[0064] Specifically, performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after multiple training actions and outputting it as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone group, and further comprising: a numerical simulation mode can be selected to implement testing and simulation of each training action of the convolutional neural network;

[0065] Wherein, obtaining various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route, further comprising: the number of flight starting points in the flight starting point set of the target flight route is equal to the number of drones in the current drone group, and each flight starting point of the target flight route is used as each starting point position corresponding to each drone of the current drone group that executes the flight of the target flight route;

[0066] And wherein, various route data of the target flight route that is about to be completed by the current drone group composed of multiple drones are obtained, and the various route data of the target flight route are the three-dimensional coordinate data of the flight starting point set of the target flight route, the three-dimensional coordinate data of the flight terminal set of the target flight route, and multiple three-dimensional coordinate data corresponding to multiple obstacle points between the flight starting point set and the flight terminal set of the target flight route. It also includes: the number of flight terminal points in the flight terminal set of the target flight route is equal to the number of drones in the current drone group, and each flight terminal point of the target flight route is used as the terminal position corresponding to each drone of the current drone group that executes the target flight route flight.

[0067] Second embodiment

[0068] Figure 3 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to a second embodiment of the present invention.

[0069] like Figure 3 As shown, compared with Figure 2 , before step S24, the method further includes:

[0070] Step S31: using the remote controller of the current drone group to simulate the various flight paths used by the current drone group to complete the target flight route, and obtaining the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by each flight path;

[0071] For example, the remote control device of the current drone swarm is used to simulate the various flight paths used by the current drone swarm to complete the target flight route, and obtain the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by each flight path, including: the control method adopted by the remote control device of the current drone swarm is a centralized control method, that is, a main control computer or controller is used to manage the entire drone swarm.

[0072] Third embodiment

[0073] Figure 4 The present invention is a flowchart showing a method for rapid planning of paths for multiple UAV area coverage according to a third embodiment of the present invention.

[0074] like Figure 4 As shown, compared with Figure 2 After step S25, the method further includes:

[0075] Step S41: Displaying in real time on the display screen of the remote controller of the current drone group the planned flight path of the current drone group to complete the target flight route, and displaying in real time the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by the planned flight path;

[0076] Specifically, the planned flight path of the current drone swarm to complete the target flight route is displayed in real time on the display screen of the remote control device of the current drone swarm, and the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by the planned flight path is displayed in real time, including: the display screen of the remote control device of the current drone swarm is an LED display screen or an LCD display screen.

[0077] Fourth embodiment

[0078] Figure 5 The present invention is a flowchart showing the steps of a method for rapid planning of paths for multi-UAV area coverage according to the fourth embodiment of the present invention.

[0079] like Figure 5 As shown, compared with Figure 2 , after step S25, the method further includes:

[0080] Step S51: wirelessly sending the planned flight path of the current drone group to complete the target flight route to a remote big data flight management network element via a wireless communication link;

[0081] For example, wirelessly sending the planned flight path of the current drone swarm to complete the target flight route to a remote big data flight management network element via a wireless communication link includes: the wireless communication link is a time division duplex communication link or a frequency division duplex communication link.

[0082] Fifth embodiment

[0083] Figure 6 The present invention is a flowchart showing a method for quickly planning paths for multiple UAV area coverage according to the fifth embodiment of the present invention.

[0084] like Figure 6 As shown, compared with Figure 2 After step S23, the method further includes:

[0085] Step S61: completing the model storage of the intelligent path planning model by storing various model parameters of the intelligent path planning model;

[0086] Specifically, completing the model storage of the intelligent path planning model by storing various model parameters of the intelligent path planning model includes: selecting to use a TF storage chip or FLASH flash memory to complete the model storage of the intelligent path planning model by storing various model parameters of the intelligent path planning model.

[0087] Next, various embodiments of the present invention will be further described.

[0088] In each of the above embodiments, optionally, in the method for rapid planning of multi-UAV area coverage paths:

[0089] Input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and run the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model;

[0090] For example, a parallel communication interface may be used to input three-dimensional coordinate data corresponding to each spatial point of a flight area covered by a certain flight path used by the current drone group to complete a target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model in parallel;

[0091] The flight path corresponding to the minimum value of each time duration corresponding to each flight path used by the current drone group to complete the target flight route is sent to the remote control device of the current drone group as the planned flight path for the current drone group to complete the target flight route.

[0092] In each of the above embodiments, optionally, in the method for rapid planning of multi-UAV area coverage paths:

[0093] Acquire various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route;

[0094] For example, obtaining various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route, respectively, includes: using a three-dimensional space coordinate system to realize a numerical representation of a single three-dimensional coordinate data corresponding to each obstacle point;

[0095] Among them, obtaining the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route as multiple flight configuration information outputs of the current drone group include: when the current drone group executes the target flight route, the flight speeds of each drone are the same and equal to the set flight speed, and the flight starting point in the flight starting point set of the target flight route is used as the first spatial point, and the flight end point in the flight end point set of the target flight route is used as the second spatial point, the shortest distance between any first spatial point and any second spatial point is obtained to obtain multiple shortest distances, and the maximum value of the multiple shortest distances is used as the longest distance;

[0096] Among them, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group are input into the intelligent path planning model, and the intelligent path planning model is run to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model. It also includes: any flight path used by the current drone group to complete the target flight route is extracted and configured to the current drone group before the current drone group completes the target flight route, and before the intelligent path planning model is run to output the corresponding time, the current drone group has not actually flown.

[0097] And in each of the above embodiments, optionally, in the method for rapid planning of multi-UAV area coverage paths:

[0098] Performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after multiple training actions and outputting the convolutional neural network as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm comprises: using a content mapping function to represent a content mapping relationship in which the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm;

[0099] Specifically, the content mapping relationship of using a content mapping function to represent the monotonically positive correlation between the number of training actions that the convolutional neural network has undergone and the number of drones in the current drone group includes: a programmable logic device can be used to implement testing and simulation of a data processing process of the content mapping function;

[0100] Among them, the content mapping relationship in which the number of training actions of the convolutional neural network and the number of drones in the current drone swarm are monotonically positively correlated using a content mapping function includes: in the content mapping function, the number of drones in the current drone swarm is the input content of the content mapping function, and the number of training actions of the convolutional neural network corresponding to the number of drones in the current drone swarm is the output content of the content mapping function.

[0101] In addition, in a method for rapid planning of multi-UAV area coverage paths according to the present invention:

[0102] Inputting the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and running the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model, including: using a numerical simulation mode to implement the testing and simulation of the data processing process of inputting the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and running the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model;

[0103] For example, a numerical simulation mode is used to input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and the intelligent path planning model is run to obtain the test and simulation of the data processing process of the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model, including: using a MATLAB toolbox to input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and the intelligent path planning model is run to obtain the test and simulation of the data processing process of the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model;

[0104] And wherein, inputting the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various line data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and running the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model also includes: performing numerical normalization processing on the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various line data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group, and then inputting them into the intelligent path planning model, and the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model is a representation after numerical normalization processing;

[0105] And by way of example, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone swarm to complete the target flight route, the various route data of the target flight route that the current drone swarm is about to complete, and the multiple flight configuration information of the current drone swarm are numerically normalized and then input into the intelligent path planning model. The duration used by the current drone swarm to complete the target flight route using the certain flight path output by the intelligent path planning model is a representation after numerical normalization, including: the numerical normalization processing is an octal numerical conversion processing.

[0106] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms under the guidance of the present invention, all of which are within the protection of the present invention.

Claims

1. A method for rapid planning of multi-UAV regional coverage paths, characterized in that: The method comprises: Acquire various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route; Obtain the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route, so as to output multiple flight configuration information of the current drone group; Performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after the multiple training actions and outputting it as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm; Input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and run the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model; Send the flight path corresponding to the minimum value of each duration corresponding to each flight path used by the current drone group to complete the target flight route to the remote controller of the current drone group as the planned flight path for the current drone group to complete the target flight route; Among them, in each training action performed on the convolutional neural network, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete a certain past flight route, the route data of the past flight route that the current drone group has completed, the set flight speed when the current drone group executes the past flight route, the radius of the minimum sphere that can accommodate the flight starting point set and the flight end point set of the past flight route in space, the number of drones in the current drone group, and the longest distance from the flight starting point set of the past flight route to the flight end point set of the past flight route are used as various input data of the convolutional neural network, and the actual time used by the current drone group to complete the past flight route using the certain flight path is used as the single output data of the convolutional neural network; Among them, the input content and output content of the intelligent path planning model are both representations after numerical normalization processing.

2. The method for rapid planning of multi-UAV regional coverage paths according to claim 1, characterized in that: Input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route to be completed by the current drone group, and the multiple flight configuration information of the current drone group into the intelligent path planning model, and run the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model, including: each spatial point of the flight area covered by any flight path used by the current drone group to complete the target flight route does not include multiple obstacle points between the flight starting point set and the flight end point set of the target flight route, so as to avoid the multiple obstacle points; Among them, various route data of the target flight route that is about to be completed by the current drone group composed of multiple drones are obtained, and the various route data of the target flight route are the three-dimensional coordinate data of the flight starting point set of the target flight route, the three-dimensional coordinate data of the flight terminal set of the target flight route, and multiple sets of three-dimensional coordinate data corresponding to multiple obstacle points between the flight starting point set and the flight terminal set of the target flight route respectively include: the multiple obstacle points between the flight starting point set and the flight terminal set of the target flight route are multiple obstacle points existing in the smallest sphere that can spatially accommodate the flight starting point set and the flight terminal set of the target flight route.

3. The method for rapid planning of multi-UAV regional coverage paths as claimed in claim 2, characterized in that: Before inputting the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and running the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model, the method further includes: The remote controller of the current drone group is used to simulate the various flight paths used by the current drone group to complete the target flight route, and the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by each flight path are obtained.

4. The method for rapid planning of multi-UAV regional coverage paths as claimed in claim 2, characterized in that: After sending the flight path corresponding to the minimum value of each duration corresponding to each flight path used by the current drone group to complete the target flight route as the planned flight path for the current drone group to complete the target flight route to the remote control device of the current drone group, the method further includes: The display screen of the remote controller of the current drone group displays in real time the planned flight path of the current drone group to complete the target flight route, and displays in real time the three-dimensional coordinate data corresponding to each spatial point in the flight area covered by the planned flight path.

5. The method for rapid planning of multi-UAV regional coverage paths as claimed in claim 2, characterized in that: After sending the flight path corresponding to the minimum value of each duration corresponding to each flight path used by the current drone group to complete the target flight route as the planned flight path for the current drone group to complete the target flight route to the remote control device of the current drone group, the method further includes: The planned flight path of the current drone swarm to complete the target flight route is wirelessly sent to the remote big data flight management network element through a wireless communication link.

6. The method for rapid planning of multi-UAV regional coverage paths as claimed in claim 2, characterized in that: After performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after multiple training actions and outputting it as an intelligent path planning model, and the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone group, the method further includes: The model storage of the intelligent path planning model is completed by storing various model parameters of the intelligent path planning model.

7. The method for rapid planning of multi-UAV area coverage paths as claimed in claim 6, characterized in that: Input the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, various route data of the target flight route to be completed by the current drone group, and multiple flight configuration information of the current drone group into the intelligent path planning model, and run the intelligent path planning model to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model; Send the flight path corresponding to the minimum value of each duration corresponding to each flight path used by the current drone group to complete the target flight route to the remote controller of the current drone group as the planned flight path for the current drone group to complete the target flight route; Wherein, obtaining various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route, further comprising: the number of flight starting points in the flight starting point set of the target flight route is equal to the number of drones in the current drone group, and each flight starting point of the target flight route is used as each starting point position corresponding to each drone of the current drone group that executes the flight of the target flight route; Among them, various route data of the target flight route that is about to be completed by the current drone swarm composed of multiple drones are obtained, and the various route data of the target flight route are the three-dimensional coordinate data of the flight starting point set of the target flight route, the three-dimensional coordinate data of the flight terminal set of the target flight route, and multiple three-dimensional coordinate data corresponding to multiple obstacle points between the flight starting point set and the flight terminal set of the target flight route. It also includes: the number of flight terminal points in the flight terminal set of the target flight route is equal to the number of drones in the current drone swarm, and each flight terminal point of the target flight route is used as each terminal position corresponding to each drone of the current drone swarm that executes the target flight route flight.

8. The method for rapid planning of multi-UAV regional coverage paths according to claim 6, characterized in that: Acquire various route data of a target flight route that is about to be completed by a current drone group composed of multiple drones, wherein the various route data of the target flight route are three-dimensional coordinate data of a flight starting point set of the target flight route, three-dimensional coordinate data of a flight end point set of the target flight route, and a plurality of three-dimensional coordinate data corresponding to a plurality of obstacle points between the flight starting point set and the flight end point set of the target flight route; Among them, obtaining the set flight speed when the current drone group executes the target flight route, the radius of the smallest sphere that can spatially accommodate the flight starting point set and the flight end point set of the target flight route, the number of drones in the current drone group, and the longest distance from the flight starting point set of the target flight route to the flight end point set of the target flight route as multiple flight configuration information outputs of the current drone group include: when the current drone group executes the target flight route, the flight speeds of each drone are the same and equal to the set flight speed, and the flight starting point in the flight starting point set of the target flight route is used as the first spatial point, and the flight end point in the flight end point set of the target flight route is used as the second spatial point, the shortest distance between any first spatial point and any second spatial point is obtained to obtain multiple shortest distances, and the maximum value of the multiple shortest distances is used as the longest distance; Among them, the three-dimensional coordinate data corresponding to each spatial point of the flight area covered by a certain flight path used by the current drone group to complete the target flight route, the various route data of the target flight route that the current drone group is about to complete, and the multiple flight configuration information of the current drone group are input into the intelligent path planning model, and the intelligent path planning model is run to obtain the time taken by the current drone group to complete the target flight route using the certain flight path output by the intelligent path planning model. It also includes: any flight path used by the current drone group to complete the target flight route is extracted and configured to the current drone group before the current drone group completes the target flight route, and before the intelligent path planning model is run to output the corresponding time, the current drone group has not actually flown.

9. The method for rapid planning of multi-UAV area coverage paths according to claim 6, characterized in that: Performing multiple training actions on the convolutional neural network to obtain a convolutional neural network after multiple training actions and outputting the convolutional neural network as an intelligent path planning model, wherein the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm comprises: using a content mapping function to represent a content mapping relationship in which the number of training actions of the convolutional neural network is monotonically positively correlated with the number of drones in the current drone swarm; Among them, the content mapping relationship in which the number of training actions of the convolutional neural network and the number of drones in the current drone swarm are monotonically positively correlated using a content mapping function includes: in the content mapping function, the number of drones in the current drone swarm is the input content of the content mapping function, and the number of training actions of the convolutional neural network corresponding to the number of drones in the current drone swarm is the output content of the content mapping function.

Images