Unmanned aerial vehicle route planning method for low-altitude airspace unmanned aerial vehicle operation

Abstract

The invention is suitable for the technical field of unmanned aerial vehicle route planning, and provides an unmanned aerial vehicle route planning method for low-altitude airspace unmanned aerial vehicle operation, which comprises the following steps of: constructing dynamic route networks at different moments based on three-dimensional GIS (Geographic Information System) gridding; evaluating the comprehensive risk of any grid at different moments based on the multi-source heterogeneous data; multiple uncertainties in a flight environment are considered, and a double-layer planning robust optimization model of unmanned aerial vehicle route planning is established according to multiple origin-destination points and time windows of unmanned aerial vehicle tasks; and designing a two-stage distributed algorithm of the bilayer planning robust optimization model for solving the route planning of the unmanned aerial vehicle. On the basis of three-dimensional GIS gridding, factors such as obstacle distribution, wind shear, a dynamic restriction area, radio interference and unmanned aerial vehicle performance are comprehensively considered, targets such as safety, efficiency and green are considered according to origin-destination points and time windows of a plurality of unmanned aerial vehicle tasks, a plurality of unmanned aerial vehicle paths are generated, and the safe and efficient level of low-altitude airspace operation is assisted.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV) route planning, and in particular relates to a UAV route planning method for UAV operation in low-altitude airspace. Background Art

[0002] Unmanned aerial vehicles (UAVs) have the characteristics of high flexibility, strong maneuverability, low safety risk coefficient, and low cost. They are widely used in search and patrol, reconnaissance and surveillance, disaster relief, logistics distribution, power inspection, agricultural irrigation and other fields. UAV route planning is the core part of the flight mission planning system. It means that the UAV continuously adjusts the flight altitude to find the best flight trajectory that matches the morphological changes of the three-dimensional restricted flight area under the premise of comprehensively considering the mission start and end points and time, fuel consumption, threats, and flight areas. It is more complicated than traditional vehicle path planning and has attracted widespread attention. Therefore, it is urgent to use big data and artificial intelligence methods to automatically generate UAV route planning for safe operation in low-altitude airspace to assist in the daily management of UAVs.

[0003] At present, the existing technologies disclose the gridding of real three-dimensional space and the use of operational optimization ideas to study the generation of drone tracks under different objectives (such as: highest task completion efficiency, shortest task completion time, minimum total operating cost, shortest flight distance, and minimum energy consumption) and constraints (such as: drone performance, obstacle and threat distribution, energy consumption, and whether there is conflicting flight). They all separate drone track planning and running time. Only Yan Shaoqiang et al. discussed the coordinated optimization of drone track planning and running time considering complex terrain. Moreover, most of the existing technologies make drone track plans based on static environments, and rarely involve dynamic environments. Regarding the real-time adjustment of UAV trajectories, only Wang Hanyang and Li Bo et al. discussed the problem of online trajectory planning of UAVs under dynamic threats; in addition, there are few technical solutions that integrate UAV delivery routes with trajectory planning. For example, Wei Ming et al. proposed a model for UAV delivery routes and trajectory collaborative planning, and Wang Fei et al. proposed a two-layer task planning method for multi-UAV logistics distribution. The solution algorithms are generally divided into heuristic algorithms (such as improved fast search random tree algorithm, sparse A* algorithm, etc.), swarm intelligence algorithms (such as particle swarm optimization algorithm, whale optimization algorithm, etc.), and deep learning (such as deep Q network algorithm, pointer network, etc.).

[0004] In summary, the technical solutions of the prior art have the following deficiencies: the existing UAV route planning is based on a static environment; it does not involve multiple risk factors; and it does not consider uncertain distributed robust optimization. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide a method for UAV route planning for UAV operation in low-altitude airspace, aiming to solve the problems raised in the above-mentioned background technology.

[0006] The embodiment of the present invention is implemented as follows: a method for UAV route planning for UAV operation in low-altitude airspace, comprising the following steps:

[0007] Construct dynamic route networks at different times based on 3D GIS gridding;

[0008] Evaluate the comprehensive risk of any grid at different times based on multi-source heterogeneous data;

[0009] Considering the multiple uncertainties in the flight environment, a two-level planning robust optimization model for UAV route planning is established based on the multiple starting and ending points and time windows of the UAV mission.

[0010] A two-stage distributed algorithm is designed to solve the bi-level programming robust optimization model for UAV path planning.

[0011] Preferably, the step of constructing a dynamic route network at different times based on three-dimensional GIS gridding specifically includes:

[0012] Grid the monitored low-altitude airspace into three dimensions and calibrate the coordinates of all grids;

[0013] Determine the traffic status of all grids at different times;

[0014] Calculate whether there is a feasible path between any adjacent grids at different times.

[0015] Preferably, the step of determining the traffic status of all grids at different times is divided into two cases:

[0016] Considering the obstacle distribution and radio interference, determine the static traffic status of some affected grids at all times;

[0017] Taking into account wind shear, dynamic restricted areas, and different safety thresholds, the static traffic status of some affected grids at different times is determined.

[0018] Preferably, the step of calculating whether there is a feasible path between any adjacent grids at different times is divided into two cases:

[0019] At a certain moment, check that the traffic status of any two adjacent grids is not prohibited from flying, and initially generate a feasible path;

[0020] Considering the performance of the UAV, check whether all feasible paths meet the requirements and remove some infeasible paths.

[0021] Preferably, the step of evaluating the comprehensive risk of any grid at different times based on multi-source heterogeneous data specifically includes:

[0022] Based on multi-source heterogeneous data, the wind shear, radio interference, and ground crash risk values ​​of each grid at any time are obtained, and their risk scores are evaluated;

[0023] The weight coefficient is estimated, and the risk score is linearly weighted to obtain the comprehensive risk score of any grid at different times.

[0024] Preferably, the step of establishing a two-level planning robust optimization model for UAV route planning based on multiple starting and ending points and time windows of UAV missions considering multiple uncertainties in the flight environment specifically includes:

[0025] Identify the decision variables of the upper and lower models;

[0026] Clarify the objective functions of the upper and lower models;

[0027] Clarify the constraints of the upper and lower models;

[0028] Based on probability distribution, various multiple uncertainties in the flight environment are characterized and converted into a linear model.

[0029] Preferably, the steps of designing a two-stage distributed algorithm for solving a two-level programming robust optimization model for UAV route planning specifically include:

[0030] Data preparation, setting parameters, including multiple UAV mission start and end points and time windows, airspace three-dimensional gridding, and other parameters;

[0031] In the first stage, according to the starting and ending points of multiple drone missions, the intelligent algorithm is deployed on different servers and runs independently to calculate several alternative feasible paths for each drone's starting and ending points, and calculate their mileage and total risk scores;

[0032] Combine several alternative feasible paths from the starting and ending points of each drone to obtain multiple complete paths from the starting point to the end point of all drones, and sort them from small to large according to mileage distance;

[0033] Each complete path of each UAV is traversed step by step, and the operation schedule plan of the complete path of the UAV is generated based on the greedy strategy using the polynomial heuristic algorithm in the second stage;

[0034] Check whether the operation schedule of all drones meets the time window requirements, and step by step check whether these drones meet the safety factors when passing through the grid at each time, and whether there is a collision risk between different drones in the same grid. If all are satisfied, output the optimal complete path and operation schedule of multiple drones; otherwise, repeat the previous step.

[0035] Preferably, the other parameters include a maximum route length and a risk threshold.

[0036] An embodiment of the present invention provides a method for UAV route planning for UAV operation in low-altitude airspace. On the basis of three-dimensional GIS gridding, it comprehensively considers factors such as obstacle distribution, wind shear, dynamic restricted areas, radio interference, and UAV performance, establishes a two-layer planning distributed robust optimization model for UAV route planning and a two-stage algorithm thereof, and generates several UAV paths (starting from the starting point, passing through several adjacent grids, and reaching the end point) according to the starting and ending points and time windows of multiple UAV tasks, taking into account goals such as safety, efficiency, and greenness, and determines the flight speeds of different UAVs in adjacent grids and the waiting time of each grid, thereby assisting the safety and efficiency level of low-altitude airspace operation and providing data support for UAV monitoring and command. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flowchart of a method for UAV route planning for UAV operation in low-altitude airspace provided by an embodiment of the present invention;

[0038] Figure 2 A flowchart of the steps of constructing a dynamic route network at different times based on three-dimensional GIS gridding provided by an embodiment of the present invention;

[0039] Figure 3 A flowchart of the steps of evaluating the comprehensive risk of any grid at different times based on multi-source heterogeneous data provided by an embodiment of the present invention;

[0040] Figure 4 A flowchart of the steps of establishing a bi-level planning distributed robust optimization model for UAV route planning based on multiple starting and ending points and time windows of UAV missions, taking into account various multiple uncertainties in the flight environment provided by an embodiment of the present invention;

[0041] Figure 5 A flowchart of the two-stage distributed algorithm steps for designing and solving a distributed robust optimization model for a two-layer planning of UAV route planning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0043] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0044] like Figure 1As shown, it is a flow chart of a method for UAV route planning for UAV operation in low-altitude airspace provided by an embodiment of the present invention. According to the starting and ending points and time windows of multiple UAV tasks, multiple uncertainties in the flight environment are considered, and safety, efficiency, and green goals are taken into account. Multiple UAV paths are generated, and the delayed departure time of different UAVs and the waiting time in each grid are determined. Specifically, the following steps are included:

[0045] S1. Construct dynamic route network at different times based on 3D GIS gridding;

[0046] S2, evaluate the comprehensive risk of any grid at different times based on multi-source heterogeneous data;

[0047] S3. Considering various uncertainties in the flight environment, a two-level distributed robust optimization model for UAV route planning is established based on multiple starting and ending points and time windows of UAV missions.

[0048] S4. Design a two-stage distributed algorithm to solve the distributed robust optimization model of bi-level planning for UAV path planning.

[0049] like Figure 2 As shown, as a preferred embodiment of the present invention, the step S1 constructs a dynamic route network at different times based on three-dimensional GIS gridding, specifically including:

[0050] S11. Grid the monitored low-altitude airspace into three dimensions and calibrate the coordinates of all grids;

[0051] S12. Determine the traffic status of all grids at different times, which is mainly divided into two situations:

[0052] First, considering the obstacle distribution and radio interference, the static traffic state of some affected grids at all times is determined;

[0053] Secondly, considering wind shear, dynamic restricted areas, and different safety thresholds, the static traffic status of some affected grids at different times is determined;

[0054] S13. Calculate whether there is a feasible path between any adjacent grids at different times. There are mainly two cases:

[0055] First, at a certain moment, check that the traffic status of any two adjacent grids is not prohibited from flying, and initially generate a feasible path;

[0056] Secondly, consider the performance of the UAV, such as climbing, turning, etc., check whether all feasible paths meet the requirements, and remove some infeasible paths.

[0057] This step constructs a dynamic route network at different times based on three-dimensional GIS gridding, calculates the traffic status of any grid under different safety thresholds, and whether there is a feasible path between adjacent grids.

[0058] like Figure 3 As shown, as a preferred embodiment of the present invention, step S2 is a step of evaluating the comprehensive risk of any grid at different times based on multi-source heterogeneous data, specifically comprising:

[0059] S21. Obtain the values ​​of wind shear, radio interference, and ground crash risk for each grid at any time based on multi-source heterogeneous data, and evaluate their risk scores according to industry safety standards;

[0060] S22, estimated weight coefficient;

[0061] S32. Linearly weight the risk score evaluated in S21 to obtain the comprehensive risk score of any grid at different times.

[0062] like Figure 4 As shown, as a preferred embodiment of the present invention, the step S3 considers multiple uncertainties in the flight environment, and establishes a two-level planning robust optimization model for UAV route planning according to multiple starting and ending points and time windows of the UAV mission, specifically including:

[0063] S31. Clarify the decision variables of the upper and lower models;

[0064] S32. Clarify the objective functions of the upper and lower models;

[0065] S33. Clarify the constraints of the upper and lower models;

[0066] S34. Characterize various multiple uncertainties in the flight environment based on probability distribution and transform them into a linear model.

[0067] This step considers the multiple uncertainties in the flight environment and establishes a two-level distributed robust optimization model for UAV route planning. Based on the multiple starting and ending points and time windows of the UAV mission, it takes into account safety, efficiency, and green goals, generates feasible paths for these UAVs, and determines their delayed departure time and waiting time in each grid to avoid conflict and interference between different UAVs in the same network.

[0068] In the upper model, considering multiple risk uncertainties such as crashing to the ground and radio, a distributed robust optimization model for UAV path planning is established. Based on multiple starting and ending points of UAV missions, considering the risk spatial distribution of the route network, the performance of UAVs and other realistic constraints, different feasible paths for UAVs are generated to avoid high-risk areas.

[0069] In the lower-level model, multiple uncertain factors such as wind shear are taken into account to establish a distributed robust optimization model for the UAV operation schedule. According to the multiple start-end time windows of the UAV mission, the delayed departure time of the UAV and the waiting time in each grid, as well as the flight speed between adjacent networks, are determined to avoid conflicts and interferences between different UAVs in the same network.

[0070] like Figure 5 As shown, as a preferred embodiment of the present invention, the step S4 designs a two-stage distributed algorithm for solving the two-level planning robust optimization model for UAV route planning, specifically including:

[0071] S41, data preparation, setting parameters, including multiple UAV mission starting and ending points and time windows, airspace three-dimensional gridding, and other parameters, such as: maximum route length, risk threshold, etc.;

[0072] S42. In the first stage, according to the starting and ending points of multiple drone missions, the intelligent algorithm is deployed on different servers and runs independently, and several alternative feasible paths for each drone starting and ending point are calculated, and their mileage and total risk scores are calculated;

[0073] S43, combining several alternative feasible paths of the starting and ending points of each drone, obtaining multiple complete paths of all drones from the starting point to the end point, and sorting them from small to large according to mileage distance;

[0074] S44, gradually traversing each complete path of each drone, and using the polynomial heuristic algorithm of the second stage to generate an operation schedule plan for the complete path of the drone based on a greedy strategy;

[0075] S45. Check whether the operating schedule of all drones meets the time window requirements, and gradually check whether these drones meet the safety factors such as dynamic restricted areas and wind shear when passing through the grids at each time, and whether there is a risk of collision between different drones in the same grid. If all the above factors are met, go to step S46; otherwise, go to step S44.

[0076] S46. Output the optimal complete paths and operation schedule plans of multiple drones.

[0077] In the first stage, this step uses a distributed parallel algorithm to search for feasible paths for different drones, and in the second stage, a polynomial heuristic algorithm is used to generate a drone operation schedule, including: determining the delayed departure time of the drone and the waiting time in each grid, as well as the flight speed between adjacent networks to avoid conflict and interference between different drones in the same network.

[0078] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for UAV route planning for UAV operation in low-altitude airspace, characterized in that: The following steps are involved: Construct dynamic route networks at different times based on 3D GIS gridding; Evaluate the comprehensive risk of any grid at different times based on multi-source heterogeneous data; Considering the multiple uncertainties in the flight environment, a two-level planning robust optimization model for UAV route planning is established based on the multiple starting and ending points and time windows of the UAV mission. A two-stage distributed algorithm is designed to solve the bi-level programming robust optimization model for UAV path planning.

2. The method for UAV route planning for UAV operation in low altitude airspace according to claim 1 is characterized in that: The steps of constructing a dynamic route network at different times based on three-dimensional GIS gridding specifically include: Grid the monitored low-altitude airspace into three dimensions and calibrate the coordinates of all grids; Determine the traffic status of all grids at different times; Calculate whether there is a feasible path between any adjacent grids at different times.

3. The method for UAV route planning for UAV operation in low altitude airspace according to claim 2 is characterized in that: The step of determining the traffic status of all grids at different times is divided into two cases: Considering the obstacle distribution and radio interference, determine the static traffic status of some affected grids at all times; Taking into account wind shear, dynamic restricted areas, and different safety thresholds, the static traffic status of some affected grids at different times is determined.

4. The method for UAV route planning for UAV operation in low altitude airspace according to claim 2 is characterized in that: The step of calculating whether there is a feasible path between any adjacent grids at different times is divided into two cases: At a certain moment, check that the traffic status of any two adjacent grids is not prohibited from flying, and initially generate a feasible path; Considering the performance of the UAV, check whether all feasible paths meet the requirements and remove some infeasible paths.

5. The method for UAV route planning for UAV operation in low altitude airspace according to claim 1, characterized in that: The step of evaluating the comprehensive risk of any grid at different times based on multi-source heterogeneous data specifically includes: Based on multi-source heterogeneous data, the wind shear, radio interference, and ground crash risk values ​​of each grid at any time are obtained, and their risk scores are evaluated; The weight coefficient is estimated, and the risk score is linearly weighted to obtain the comprehensive risk score of any grid at different times.

6. The method for UAV route planning for UAV operation in low altitude airspace according to claim 1, characterized in that: The steps of establishing a two-level planning robust optimization model for UAV route planning based on multiple starting and ending points and time windows of UAV missions considering multiple uncertainties in the flight environment specifically include: Identify the decision variables of the upper and lower models; Clarify the objective functions of the upper and lower models; Clarify the constraints of the upper and lower models; Based on probability distribution, various multiple uncertainties in the flight environment are characterized and converted into a linear model.

7. The method for UAV route planning for UAV operation in low altitude airspace according to claim 1, characterized in that: The steps of designing a two-stage distributed algorithm for solving a bi-level programming robust optimization model for UAV route planning specifically include: Data preparation, setting parameters, including multiple UAV mission start and end points and time windows, airspace three-dimensional gridding, and other parameters; In the first stage, according to the starting and ending points of multiple drone missions, the intelligent algorithm is deployed on different servers and runs independently to calculate several alternative feasible paths for each drone's starting and ending points, and calculate their mileage and total risk scores; Combine several alternative feasible paths from the starting and ending points of each drone to obtain multiple complete paths from the starting point to the end point of all drones, and sort them from small to large according to mileage distance; Each complete path of each UAV is traversed step by step, and the operation schedule of the complete path of the UAV is generated based on the greedy strategy using the polynomial heuristic algorithm in the second stage; Check whether the operation schedule of all drones meets the time window requirements, and step by step check whether these drones meet the safety factors when passing through the grid at each time, and whether there is a collision risk between different drones in the same grid. If all are satisfied, output the optimal complete path and operation schedule of multiple drones; otherwise, repeat the previous step.

8. The method for UAV route planning for UAV operation in low altitude airspace according to claim 7, characterized in that: The other parameters include maximum route length and risk threshold.

Images