Path planning method based on GVD and improved Theta-star algorithm

Abstract

The invention discloses a path planning method based on GVD and an improved Theta-star algorithm. The method comprises the steps that a building projection plane graph in a corresponding area is acquired; determining the resolution of the grid map according to the area size and range, and converting the building projection plane graph into the grid map; generating a recommended flight route in the unmanned aerial vehicle region by using a generalized Voronoi diagram method, and updating the recommended flight route to the grid map; calculating the cost of different grids through a certain mathematical function according to the distribution of buildings and recommended routes in the regional space, and generating a non-uniform cost grid map; and traffic bottlenecks and accident hidden dangers possibly existing in the generated flight recommendation route are calculated and analyzed through intermediary centrality. The method improves the low-altitude flight path planning speed and efficiency of the unmanned aerial vehicle, balances the flight safety and speed of the unmanned aerial vehicle, provides data support for low-altitude airspace management, and is suitable for large-scale application of urban unmanned aerial vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and in particular to a path planning method based on GVD and an improved Theta-star algorithm. Background Art

[0002] Drone path planning involves finding the optimal or feasible path from a starting point to a destination that meets the drone's performance requirements under specific constraints. Existing drone path planning technologies fall into two categories: research based on numerical algorithms, such as biomimetic algorithms or particle swarm algorithms; and research based on graphical algorithms, such as Voronoi diagrams and Laguerre diagrams. Existing technologies include research based on biomimetic algorithms like ant colonies and genetic algorithms, drone path planning based on various algorithms, such as gravitational search algorithms and particle swarm algorithms, and real-time static and dynamic obstacle avoidance path planning for various drones.

[0003] With the large-scale application of drones in urban airspace, their low-altitude flight path planning faces many challenges. Traditional path planning methods have difficulty balancing flight safety and speed when dealing with complex urban environments, and the planning efficiency is low. In urban environments, buildings are complexly distributed and there are many obstacles. At the same time, high requirements are placed on the timeliness and safety of drone flights. Existing path planning algorithms, such as the traditional Theta-star algorithm (Theta* algorithm), have problems such as insufficient consideration of environmental complexity and slow path planning when dealing with such complex environments, and cannot meet the needs of large-scale application of drones in cities. Therefore, there is an urgent need for a new path planning method to improve the speed and efficiency of drone low-altitude flight path planning, balance the issues of flight safety and speed, and provide effective data support for low-altitude airspace management. Summary of the Invention

[0004] The purpose of the present invention is to overcome the defects existing in the above-mentioned background technology. The present invention proposes a path planning method based on GVD and an improved Theta-star algorithm. According to the actual situation of UAVs flying at low altitude in complex areas such as cities, a generalized Voronoi diagram is used to construct a flight recommendation route network. The improved Theta-star algorithm is used to comprehensively consider multiple factors for path planning, thereby improving the speed and efficiency of path planning, balancing the relationship between flight safety and speed, and providing data support for low-altitude airspace management.

[0005] The present invention adopts the following technical solutions to solve the above technical problems:

[0006] The path planning method based on GVD and improved Theta-star algorithm includes the following steps:

[0007] Step S1: Obtain environmental information data within the area, including the projection distribution of buildings within the area, extract building outlines from satellite maps, and map the building outlines to the building outlines. Figure 2 The value is obtained and converted into a raster map;

[0008] Step S2: constructing a recommended flight route network for UAVs based on the GVD, dividing the binary image within the region into sub-regions and taking their intersections as recommended routes, and updating the recommended route network to the uniform cost grid map generated by the environmental information data. The GVD is a generalized Voronoi diagram.

[0009] Step S3: defining a non-uniform cost grid map according to a mathematical function, and combining the updated uniform cost grid map with the function to calculate the different cost values ​​of each grid;

[0010] Step S4: Improve the traditional Theta-star algorithm by introducing a risk assessment function, comprehensively considering the path length, the degree of closeness to the heuristic path, and the risk cost factors to design a new valuation function;

[0011] Step S5: Using the improved Theta-star algorithm, combined with the constructed flight corridor space and the evaluation function, to perform UAV path planning;

[0012] Step S6: Analyze and recommend possible traffic bottlenecks and accident risks in the route network based on the betweenness centrality calculation results to help warn of possible accidents;

[0013] The method is applied to a UAV path planning system.

[0014] Furthermore, the step S1 specifically includes:

[0015] Step S1-1, obtaining a satellite map, extracting the main part of the plane projection of the building at a certain height layer in the area, and simplifying it to reduce the amount of calculation;

[0016] Step S1-2, then specifying the grid size according to the map size and rasterizing the map;

[0017] Step S1-3: If the grid contains obstacle pixels, it is marked as 1, otherwise it is marked as 0, thereby obtaining a grid map that can be used for subsequent operations.

[0018] Furthermore, the step S2 specifically includes:

[0019] Step S2-1: Consider relatively independent obstacles as control points. Using the Voronoi diagram principle, divide the space into several sub-regions through a series of control points. The distance from any point in each sub-region to the corresponding control point is smaller than the distance to other control points, and its boundaries constitute the edges of the Voronoi diagram.

[0020] Step S2-2: On this basis, the edge farthest from the adjacent obstacle is formed based on the generalized Voronoi diagram;

[0021] Step S2-3: Use the Brushfire algorithm to obtain the edge of the GVD under the grid map, including adding a black rectangular boundary to the grid map and treating the black part as an obstacle, marking the obstacle, calculating the distance transformation, checking the adjacent background grid labels to determine the GVD boundary, and performing a morphological dilation operation on the GVD boundary. Finally, the recommended route distribution map is obtained and updated to the grid map accordingly.

[0022] Furthermore, the step S3 specifically includes:

[0023] The cost of different grids is calculated using the attractive potential field function and the repulsive potential field function to construct a non-uniform cost grid map; the specific functions are as follows:

[0024]

[0025] Among them U att is the attractive potential field function, k att is the attraction coefficient, d(x,x goal ) is the current position x to the target point x goal The farther away from the target, the greater the gravitational potential and the greater the cost;

[0026]

[0027] Among them U rep is the repulsive potential field function, where k rep is the rejection coefficient, d(x,x obs ) is the distance from the current position to the obstacle x obs The distance d0 is the threshold distance, and only within this distance will the grid cost be affected by obstacles; the single grid cost R n The calculation formula is

[0028] R n＝ U att+ U rep .

[0029] Furthermore, the new valuation function in step S4 is:

[0030] f(n)=g(n)+h(n)+R(n)

[0031] Where g(n) represents the actual cost from the starting node to the current node n. The basis for measuring this cost depends on the specific problem scenario. h(n) is a heuristic function that represents the estimated value of the optimal path cost from the current node n to the target node. It estimates the cost required to reach the target node from the current node based on certain experience or problem characteristics. R(n) is the cost of exploring the grid where node n is located.

[0032] Furthermore, the calculation method of g(n) and h(n) in the new evaluation function is as follows:

[0033]

[0034] Where (x0, y0) is the starting point coordinate, (x n ,y n ) is the coordinate of the current node n, (x t ,y t ) are the end point coordinates.

[0035] Furthermore, the step S5 specifically includes:

[0036] Step S5-1, initialize the open list, add the starting node and its initial cost, i.e., the heuristic function value, to the open list;

[0037] Step S5-2: Take the node with the lowest cost from the open list as the current node;

[0038] Step S5-3: If the current node is the target node, backtrack to obtain the complete path according to the recorded path;

[0039] Step S5-4: If the current node is not the target node, traverse the neighboring nodes of the current node and calculate the costs of the neighboring nodes, including the actual movement cost, estimated cost, and risk assessment cost. If the neighboring node is not in the cost record or the newly calculated cost is smaller, update its cost record and parent node and add it to the open list. At the same time, according to the specific logic of the Theta-star algorithm, check whether the current node and the neighboring node can be connected by a straight line. If the conditions are met, update the parent node and cost of the current node.

[0040] Step S5-5: Repeat the above process until the target node is found or the open list is empty.

[0041] Furthermore, the step S6 specifically includes:

[0042] Step S6-1: Based on the recommended route network constructed in step S2, calculate the betweenness centrality value of each route edge:

[0043] Count the number of times all possible shortest paths pass through the route edge, and normalize it to get its betweenness centrality value. The formula is:

[0044]

[0045] Among them, σ st is the total number of shortest paths from node s to t, σ st (e) is the number of shortest paths passing through route edge e, and V is the set of network nodes;

[0046] Step S6-2: Identify key route edges based on betweenness centrality values:

[0047] Set the threshold τ, if C B (e) ≥ τ, then mark the route edge as a high-flow critical path, which may become a traffic bottleneck; combined with the non-uniform cost grid map in step S3, further analyze the risk cost near the critical path and mark potential accident risk points;

[0048] Step S6-3: Generate early warning and optimization suggestions:

[0049] For the marked critical paths, the cost value is dynamically adjusted to guide the drone to prioritize alternative paths; a visual report is output to mark high-risk areas and alternative routes for operators to refer to in decision-making.

[0050] Compared with the prior art, the present invention adopts the above technical solution and has the following beneficial effects:

[0051] (1) The path planning method based on GVD and improved Theta-star algorithm proposed in this paper, by combining the generalized Voronoi diagram and the improved Theta-star algorithm, can plan the flight path of drones more quickly and efficiently compared with traditional path planning methods, thus meeting the timeliness requirements of large-scale application of urban drones.

[0052] (2) The path planning method based on GVD and improved Theta-star algorithm proposed in this invention uses a non-uniform cost grid map to comprehensively consider factors such as obstacles and target distances, so that the planned path can ensure flight speed while maximizing flight safety and avoiding dangerous situations such as collisions with obstacles.

[0053] (3) The path planning method based on GVD and improved Theta-star algorithm proposed in this invention calculates and analyzes the betweenness centrality of the recommended route network, which can discover possible traffic bottlenecks and accident risks in advance, provide decision-making basis for low-altitude airspace management departments, help optimize airspace resource allocation, and improve the scientificity and effectiveness of airspace management. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flow chart of the method of the present invention;

[0055] Figure 2 This is a schematic diagram of the present invention dividing the space into sub-regions based on the Voronoi diagram principle;

[0056] Figure 3 A recommended route map generated by the GVD of the present invention;

[0057] Figure 4 This is the betweenness centrality heat map of the recommended route map obtained by the present invention through betweenness centrality calculation. DETAILED DESCRIPTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] Path planning method based on GVD and improved Theta-star algorithm, such as Figure 1 As shown, the following steps are included:

[0060] Step S1: Obtain environmental information data within the area, including the projection distribution of buildings within the area, extract building outlines from satellite maps, and map the building outlines to the building outlines. Figure 2 The value is obtained and converted into a raster map;

[0061] Step S2: constructing a recommended flight route network for UAVs based on the GVD, dividing the binary image within the region into sub-regions and taking their intersections as recommended routes, and updating the recommended route network to the uniform cost grid map generated by the environmental information data. The GVD is a generalized Voronoi diagram.

[0062] Step S3: defining a non-uniform cost grid map according to a mathematical function, and combining the updated uniform cost grid map with the function to calculate the different cost values ​​of each grid;

[0063] Step S4: Improve the traditional Theta-star algorithm by introducing a risk assessment function, comprehensively considering the path length, the degree of closeness to the heuristic path, and the risk cost factors to design a new valuation function;

[0064] Step S5: Using the improved Theta-star algorithm, combined with the constructed flight corridor space and the evaluation function, to perform UAV path planning;

[0065] Step S6: Analyze and recommend possible traffic bottlenecks and accident risks in the route network based on the betweenness centrality calculation results to help warn of possible accidents;

[0066] The method is applied to a UAV path planning system.

[0067] Furthermore, the step S1 specifically includes:

[0068] Step S1-1, obtaining a satellite map, extracting the main part of the plane projection of the building at a certain height layer in the area, and simplifying it to reduce the amount of calculation;

[0069] Step S1-2, then specifying the grid size according to the map size and rasterizing the map;

[0070] Step S1-3: If the grid contains obstacle pixels, it is marked as 1, otherwise it is marked as 0, thereby obtaining a grid map that can be used for subsequent operations.

[0071] Furthermore, the step S2 specifically includes:

[0072] Step S2-1: Consider relatively independent obstacles as control points. Use the Voronoi diagram principle to divide the space into several sub-regions through a series of control points. The distance from any point in each sub-region to the corresponding control point is less than the distance to other control points. Its boundaries constitute the edges of the Voronoi diagram. The schematic diagram of the formed sub-region is as follows: Figure 2 As shown;

[0073] Step S2-2: On this basis, forming the edge farthest from the adjacent obstacle based on the generalized Voronoi diagram;

[0074] Step S2-3, using the Brushfire algorithm to obtain the edge of the GVD under the grid map, including adding a black rectangular boundary to the grid map and treating the black part as an obstacle, marking the obstacle, calculating the distance transformation, checking the adjacent background grid labels to determine the GVD boundary, and performing a morphological expansion operation on the GVD boundary, and finally obtaining the recommended route distribution map, as shown in Figure 5. Figure 3 As shown, the grid map is updated accordingly.

[0075] Furthermore, the step S3 specifically includes:

[0076] The cost of different grids is calculated using the attractive potential field function and the repulsive potential field function to construct a non-uniform cost grid map; the specific functions are as follows:

[0077]

[0078] Among them U att is the attractive potential field function, where katt is the attraction coefficient, d(x,x goal ) is the current position x to the target point x goal The farther away from the target, the greater the gravitational potential and the greater the cost;

[0079]

[0080] Among them U rep is the repulsive potential field function, where k rep is the rejection coefficient, d(x,x obs ) is the distance from the current position to the obstacle x obs The distance d0 is the threshold distance, and only within this distance will the grid cost be affected by obstacles; the single grid cost R n The calculation formula is

[0081] R n＝ U att+ U rep .

[0082] Furthermore, the new valuation function in step S4 is:

[0083] f(n)=g(n)+h(n)+R(n)

[0084] Where g(n) represents the actual cost from the starting node to the current node n. The basis for measuring this cost depends on the specific problem scenario. h(n) is a heuristic function that represents the estimated value of the optimal path cost from the current node n to the target node. It estimates the cost required to reach the target node from the current node based on certain experience or problem characteristics. R(n) is the cost of exploring the grid where node n is located.

[0085] Furthermore, the calculation method of g(n) and h(n) in the new evaluation function is as follows:

[0086]

[0087] Where (x0, y0) is the starting point coordinate, (x n ,y n ) is the coordinate of the current node n, (x t ,y t ) are the end point coordinates.

[0088] Furthermore, the step S5 specifically includes:

[0089] Step S5-1, initialize the open list, add the starting node and its initial cost, i.e., the heuristic function value, to the open list;

[0090] Step S5-2: Take the node with the lowest cost from the open list as the current node;

[0091] Step S5-3: If the current node is the target node, backtrack to obtain the complete path according to the recorded path;

[0092] Step S5-4: If the current node is not the target node, traverse the neighboring nodes of the current node and calculate the costs of the neighboring nodes, including the actual movement cost, estimated cost, and risk assessment cost. If the neighboring node is not in the cost record or the newly calculated cost is smaller, update its cost record and parent node and add it to the open list. At the same time, according to the specific logic of the Theta-star algorithm, check whether the current node and the neighboring node can be connected by a straight line. If the conditions are met, update the parent node and cost of the current node.

[0093] Step S5-5: Repeat the above process until the target node is found or the open list is empty.

[0094] Furthermore, the step S6 specifically includes:

[0095] Step S6-1: Based on the recommended route network constructed in step S2, calculate the betweenness centrality value of each route edge:

[0096] Count the number of times all possible shortest paths pass through the route edge, and normalize it to get its betweenness centrality value. The formula is:

[0097]

[0098] Among them, σ st is the total number of shortest paths from node s to t, σ st (e) is the number of shortest paths passing through route edge e, and V is the set of network nodes;

[0099] Step S6-2: Identify key route edges based on betweenness centrality values:

[0100] Set the threshold τ, if C B (e) ≥ τ, then mark the route edge as a high-flow critical path, which may become a traffic bottleneck; combined with the non-uniform cost grid map in step S3, further analyze the risk cost near the critical path and mark potential accident risk points;

[0101] Step S6-3: Generate early warning and optimization suggestions:

[0102] For the marked critical paths, the cost value is dynamically adjusted to guide the drone to prioritize alternative paths; a visual report is output to mark high-risk areas and alternative routes for operators to make reference. The betweenness centrality heat map is as follows: Figure 4 shown.

[0103] The above is a detailed introduction to the path planning method based on GVD and the improved Theta-star algorithm provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some or all of the technical features therein can be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A path planning method based on GVD and improved Theta-star algorithm, characterized by: The method comprises the following steps: Step S1, obtaining environmental information data within the area, the data including the projection distribution of buildings in the area, extracting building outlines from a satellite map, binarizing the building outlines, and converting them into a raster map; Step S2: constructing a recommended flight route network for UAVs based on the GVD, dividing the binary image within the region into sub-regions and taking their intersections as recommended routes, and updating the recommended route network to the uniform cost grid map generated by the environmental information data. The GVD is a generalized Voronoi diagram. Step S3: defining a non-uniform cost grid map according to a mathematical function, and combining the updated uniform cost grid map with the function to calculate the different cost values ​​of each grid; Step S4: Improve the traditional Theta-star algorithm by introducing a risk assessment function, comprehensively considering the path length, the degree of closeness to the heuristic path, and the risk cost factors to design a new valuation function; Step S5: Using the improved Theta-star algorithm, combined with the constructed flight corridor space and the evaluation function, to perform UAV path planning; Step S6: Analyze and recommend possible traffic bottlenecks and accident risks in the route network based on the betweenness centrality calculation results to help warn of possible accidents; The method is applied to a UAV path planning system.

2. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The step S1 specifically includes: Step S1-1, obtaining a satellite map, extracting the main part of the plane projection of the building at a certain height layer in the area, and simplifying it to reduce the amount of calculation; Step S1-2, then specifying the grid size according to the map size and rasterizing the map; Step S1-3: If the grid contains obstacle pixels, it is marked as 1, otherwise it is marked as 0, thereby obtaining a grid map that can be used for subsequent operations.

3. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The step S2 specifically includes: Step S2-1: Consider relatively independent obstacles as control points. Using the Voronoi diagram principle, divide the space into several sub-regions through a series of control points. The distance from any point in each sub-region to the corresponding control point is smaller than the distance to other control points, and its boundaries constitute the edges of the Voronoi diagram. Step S2-2: On this basis, the edge farthest from the adjacent obstacle is formed based on the generalized Voronoi diagram; Step S2-3: Use the Brushfire algorithm to obtain the edge of the GVD under the grid map, including adding a black rectangular boundary to the grid map and treating the black part as an obstacle, marking the obstacle, calculating the distance transformation, checking the adjacent background grid labels to determine the GVD boundary, and performing a morphological dilation operation on the GVD boundary. Finally, the recommended route distribution map is obtained and updated to the grid map accordingly.

4. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The step S3 specifically includes: The cost of different grids is calculated using the attractive potential field function and the repulsive potential field function to construct a non-uniform cost grid map; the specific functions are as follows: Among them U att is the attractive potential field function, k att is the attraction coefficient, d(x,x goal ) is the current position x to the target point x goal The farther away from the target, the greater the gravitational potential and the greater the cost; Among them U rep is the repulsive potential field function, k rep is the rejection coefficient, d(x,x obs ) is the distance from the current position to the obstacle x obs The distance d0 is the threshold distance, and only within this distance will the grid cost be affected by obstacles; the single grid cost R n The calculation formula is R n＝ U att+ U rep 。 5. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The new valuation function in step S4 is: f(n)=g(n)+h(n)+R(n) Where g(n) represents the actual cost from the starting node to the current node n. The basis for measuring this cost depends on the specific problem scenario. h(n) is a heuristic function that represents the estimated value of the optimal path cost from the current node n to the target node. It estimates the cost required to reach the target node from the current node based on certain experience or problem characteristics. R(n) is the cost of exploring the grid where node n is located.

6. The path planning method based on GVD and improved Theta-star algorithm according to claim 5, characterized in that: The calculation method of g(n) and h(n) in the new evaluation function is as follows: Where (x0, y0) is the starting point coordinate, (x n ,y n ) is the coordinate of the current node n, (x t ,y t ) are the end point coordinates.

7. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The step S5 specifically includes: Step S5-1, initialize the open list, add the starting node and its initial cost, i.e., the heuristic function value, to the open list; Step S5-2: Take the node with the lowest cost from the open list as the current node; Step S5-3: If the current node is the target node, backtrack to obtain the complete path according to the recorded path; Step S5-4: If the current node is not the target node, traverse the neighboring nodes of the current node and calculate the costs of the neighboring nodes, including the actual movement cost, estimated cost, and risk assessment cost. If the neighboring node is not in the cost record or the newly calculated cost is smaller, update its cost record and parent node and add it to the open list. At the same time, according to the specific logic of the Theta-star algorithm, check whether the current node and the neighboring node can be connected by a straight line. If the conditions are met, update the parent node and cost of the current node. Step S5-5: Repeat the above process until the target node is found or the open list is empty.

8. The path planning method based on GVD and improved Theta-star algorithm according to claim 1, characterized in that: The step S6 specifically includes: Step S6-1: Based on the recommended route network constructed in step S2, calculate the betweenness centrality value of each route edge: Count the number of times all possible shortest paths pass through the route edge, and normalize it to get its betweenness centrality value. The formula is: Among them, σ st is the total number of shortest paths from node s to t, σ st (e) is the number of shortest paths passing through route edge e, and V is the set of network nodes; Step S6-2: Identify key route edges based on betweenness centrality values: Set the threshold τ, if C B (e) ≥ τ, then mark the route edge as a high-flow critical path, which may become a traffic bottleneck; combined with the non-uniform cost grid map in step S3, further analyze the risk cost near the critical path and mark potential accident risk points; Step S6-3: Generate early warning and optimization suggestions: For the marked critical paths, the cost value is dynamically adjusted to guide the drone to prioritize alternative paths; a visual report is output to mark high-risk areas and alternative routes for operators to refer to in decision-making.

Images