An unmanned aerial vehicle path planning method and system coupling a wind field and a potential field method

Abstract

The application provides an unmanned aerial vehicle path planning method and system coupling a wind field and a potential field method, and belongs to the technical field of unmanned aerial vehicle path planning, and comprises the following steps: obtaining air pressure data and environment data of a target region; constructing a two-dimensional grid map of the target region, calculating the gradient of the air pressure of a grid point based on the air pressure data to obtain a simulated wind field, initializing a sampling space based on the environment data, and establishing a potential field; performing random sampling in the sampling space, controlling the growth direction of the next expansion of a random tree based on the coupling force of the simulated wind field and the potential field, and generating a new node along the direction; performing collision checking on the generated new node; backtracking from the last sampling point around a target point to a starting point to determine a path, and performing pruning and smoothing to complete path planning. The application guides the sampling direction by coupling the wind field and the potential field, so that the consumption of flight energy is reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of unmanned aerial vehicle path planning, and particularly relates to an unmanned aerial vehicle path planning method and system coupling a wind field and an artificial potential field method. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] With the continuous development of unmanned aerial vehicle technology, its use in modern society is increasingly widespread, covering many fields such as emergency communication, traffic management, logistics transportation, forest fire prevention, data collection, etc. Path planning is one of the key technologies for unmanned aerial vehicles to achieve autonomous motion, which can help unmanned aerial vehicles plan the optimal or suboptimal path in a complex environment and guide the unmanned aerial vehicles from the starting point to the target position.

[0004] In the prior art, the Rapidly-exploring Random Tree (RRT) is a common path planning algorithm. Its advantages lie in strong adaptability to high-dimensional, complex, dynamic environments and efficient real-time computing capability, but it has certain deficiencies: its strong randomness leads to poor stability of the planned path, and the generated path is mostly a broken line, which does not conform to the kinematics and dynamics characteristics of the unmanned aerial vehicle and cannot be directly used for actual flight of the unmanned aerial vehicle. RRT* is an optimized form of RRT algorithm, which continuously optimizes the path by dynamically adjusting the tree structure while exploring the feasible path, and can gradually approach the global optimal solution. Secondly, the artificial potential field method realizes path guidance by simulating the concept of "potential field" in the physical field: the target point generates an attractive force on the unmanned aerial vehicle (pulling the unmanned aerial vehicle towards the target), and the obstacle generates a repulsive force on the unmanned aerial vehicle (blocking the unmanned aerial vehicle from approaching the obstacle); in theory, the unmanned aerial vehicle can gradually move from the "potential field high point" (starting point) to the "potential field low point" (end point); but this algorithm is prone to fall into the "local minimum" dilemma - when the attractive force and the repulsive force reach a balance, the unmanned aerial vehicle may be stationary at a non-target position and cannot continue to advance towards the end point. To make up for the defects of a single algorithm, the prior art combines the artificial potential field method with RRT*, which can use potential field guidance to sample to improve the path search efficiency of RRT*, and at the same time, use the global path finding capability of RRT* to overcome the problem of the potential field method falling into local optimum, thereby balancing the efficiency and global optimality of path planning. However, most of the existing algorithms do not consider the impact of the real environment on the unmanned aerial vehicle, such as the impact of the wind field on the energy consumption of the unmanned aerial vehicle during movement. SUMMARY

[0005] In order to overcome the above-mentioned deficiencies of the prior art, the present application proposes a UAV path planning method and system coupling a wind field and an artificial potential field method, which couples real-time data of the wind field with an artificial potential field of obstacles and the like after processing the real-time data of the wind field.

[0006] To achieve the above object, one or more embodiments of the present application provide the following technical solutions:

[0007] In a first aspect, the present application discloses a UAV path planning method coupling a wind field and an artificial potential field method, comprising:

[0008] acquiring air pressure data and environmental data of a target area;

[0009] constructing a two-dimensional grid graph for the target area, synchronously importing the air pressure data and the environmental data, calculating the gradient of the air pressure of the grid points based on the air pressure data to obtain a simulated wind field, and simultaneously initializing a sampling space based on the environmental data and establishing an artificial potential field;

[0010] starting from a starting point, performing random sampling in the sampling space, controlling the growth direction of the next expansion of a random tree based on the coupling force of the simulated wind field and the artificial potential field, and generating a new node along the direction thereof;

[0011] performing collision detection on the generated new node, if there is no collision, determining an optimal parent node and performing reconnection, and if a collision occurs, re-performing sampling until a point around the target point is sampled;

[0012] backtracking from the last sampling point around the target point to the starting point to determine a path, and performing pruning and smoothing to complete path planning.

[0013] In a second aspect, the present application discloses a UAV path planning system coupling a wind field and an artificial potential field method, comprising:

[0014] a data acquisition module configured to acquire air pressure data and environmental data of a target area;

[0015] a grid construction module configured to construct a two-dimensional grid graph for the target area, synchronously import the air pressure data and the environmental data, calculate the gradient of the air pressure of the grid points based on the air pressure data to obtain a simulated wind field, and simultaneously initialize a sampling space based on the environmental data and establish an artificial potential field;

[0016] The node growth module is configured to: start from a starting point, randomly sample in the sampling space, control a growth direction of next expansion of a random tree based on a coupling force of the simulated wind field and the artificial potential field, and generate a new node along the direction;

[0017] The node determination module is configured to: perform collision checking on the generated new node, if there is no collision, determine an optimal parent node, and perform reconnection, and if collision occurs, re-sample until a point around the target point is sampled;

[0018] The path determination module is configured to: backtrack from the last sampling point around the target point to the starting point to determine a path, and perform pruning and smoothing to complete path planning.

[0019] In a third aspect, the present application discloses an electronic device, comprising a memory and a processor, and computer instructions stored in the memory and running on the processor, when the computer instructions are run by the processor, the steps of the UAV path planning method based on coupling of wind field and artificial potential field are completed.

[0020] In a fourth aspect, the present application discloses a computer readable storage medium for storing computer instructions, when the computer instructions are executed by the processor, the steps of the UAV path planning method based on coupling of wind field and artificial potential field are completed.

[0021] Compared with the prior art, the present application has the following beneficial effects:

[0022] The present application considers the influence of wind field on the energy consumption of UAV movement in the real environment. On the basis of the RRT* algorithm, the growth direction of the random tree is guided by coupling the wind field and the artificial potential field, so that a flight path more suitable for the wind condition is planned, and the energy consumption of the UAV in the flight process is effectively reduced.

[0023] The present application makes the obtained path more consistent with the movement trajectory of the UAV through pruning and three-order Bezier curve smoothing operations.

[0024] The advantages of the additional aspects of the present application will be partially given in the following description, partially become obvious from the following description, or be known by the practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0025] The drawings accompanying the specification of the present application form a part thereof, serve to provide further understanding of the present application, and together with the description of the exemplary embodiments of the present application and their description serve to explain the present application, and do not constitute improper limitations on the present application.

[0026] Figure 1 The flowchart of the UAV path planning method based on coupling of wind field and artificial potential field described in the embodiment one of the present application.

[0027] Figure 2 A simulation diagram of the wind farm described in Embodiment One of the present application.

[0028] Figure 3 A path search process diagram described in Embodiment One of the present application.

[0029] Figure 4 A path pruning diagram described in Embodiment One of the present application.

[0030] Figure 5 A smoothing process diagram described in Embodiment One of the present application, wherein Fig. (a) is a path diagram before smoothing process, and Fig. (b) is a path diagram after smoothing process.

[0031] Figure 6 A final effect diagram described in Embodiment One of the present application. DETAILED DESCRIPTION

[0032] It should be noted that the following detailed description is merely exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the application belongs.

[0033] It is to be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application.

[0034] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0035] Embodiment One

[0036] It should be understood that the Rapidly-exploring Random Tree (RRT) is to generate a tree-like path in space by random sampling, the basic principle of which is to start from the starting point, randomly sample in free space; find the nearest tree node, extend a fixed step length towards the sampling point; after collision detection, add the new node to the tree structure; repeat until the target point is connected. Its advantages lie in strong adaptability to high-dimensional, complex and dynamic environment and high-efficiency real-time computing capability, but it has certain deficiencies: the randomness of its expansion is too strong, resulting in poor stability of the planned path, and the generated path is mostly a broken line, which does not conform to the motion dynamics characteristics of the unmanned aerial vehicle, and cannot be directly used for unmanned aerial vehicle flight. RRT* makes the path more optimal by reselecting the parent node on the basis of RRT.

[0037] In one or more embodiments, a method for unmanned aerial vehicle path planning coupling a wind field and an artificial potential field method is disclosed, as shown in Figure 1 including the following steps:

[0038] Step S1: Obtain air pressure data and environmental data for the target area.

[0039] In this embodiment, the air pressure data is the air pressure at each point within the area; the environmental data includes the location coordinates and shape information of surrounding obstacles.

[0040] Step S2: Construct a two-dimensional grid map of the target area, and simultaneously import air pressure data and environmental data. On the one hand, calculate the pressure gradient of the grid points based on the air pressure data to obtain the simulated wind field. On the other hand, initialize the sampling space and establish an artificial potential field based on the environmental data.

[0041] It should be understood that, with a two-dimensional plane as a reference, wind flow is caused by pressure gradient forces, ideally flowing from high-pressure areas to low-pressure areas. Therefore, air pressure data is used to simulate the wind field.

[0042] Step S2-1: Construct a two-dimensional grid map, synchronously import air pressure data, calculate the gradient information of air pressure at grid points, normalize the gradient vector, and obtain the direction and magnitude of the wind force. Specifically, this includes:

[0043] Ideally, the x-axis and y-axis of the plane are evenly divided. By setting a scaling factor, the grid density is adjusted (e.g., increasing the scaling factor decreases the grid spacing and increases the density), thus generating a two-dimensional grid map. By importing air pressure data, the air pressure gradient vector at each grid point in the two-dimensional plane is calculated. (Where P is air pressure,) , (These represent the rates of change of air pressure along the x and y axes, respectively), generating a regional air pressure gradient map;

[0044] Specifically, the gradient vector of each grid point in the two-dimensional plane is calculated as follows:

[0045] Calculate the gradient component between two points based on air pressure data. It represents the rate of change of air pressure along the x-axis. The gradient represents the rate of change of air pressure along the y-axis, where P1 and P2 are the air pressure values ​​at two points, and (x1, y1) and (x2, y2) are the coordinates of the two points. The magnitude of the gradient is positively correlated with the strength of the air pressure gradient force, reflecting the strength of the wind, and its direction points in the direction of the fastest increase in the function value, i.e., the opposite direction of the wind direction.

[0046] Calculate the gradient magnitude and normalize the gradient vector to obtain the normalized gradient vector.

[0047] The method for normalizing the gradient vector is as follows:

[0048] The magnitude of the gradient vector at each grid point is calculated as follows:

[0049]

[0050] wherein, is the gradient vector of each grid point, is the module length; each component of the gradient vector is divided by the module length to obtain the normalized component:

[0051]

[0052] wherein, is the normalized gradient vector.

[0053] According to the grid index, the above normalization method is applied to each point in the region grid to eliminate the difference in gradient amplitude between different grid points and ensure that the vector only retains the directional feature. Thus, a normalized gradient matrix covering the entire region is generated.

[0054] According to the basic physical law of atmospheric motion, the natural flow direction of the airflow is from the high-pressure area to the low-pressure area, that is, the wind field direction is completely opposite to the gradient direction of the pressure field. The pressure field gradient vector points to the direction of the fastest pressure increase, while the wind field moves in the direction of pressure decrease. Therefore, taking the negative of the extracted normalized gradient vector component can obtain the directional feature of the wind field at the grid point; at the same time, the normalized gradient vector itself has eliminated the absolute difference in gradient amplitude between different grid points, and its module length is 1, so the module length of the negative vector can be directly used as a standardized representation of the wind field intensity to reflect the relative strength relationship of the wind field. By combining the direction and standardized intensity information, the wind field force vector of each grid point is finally determined (F(x, y) = (Fx, Fy)), the directional feature is the direction of the wind field force vector, and the standardized intensity information is the length of the wind field force vector, realizing the quantitative description of the wind field distribution of the target region. Visualization is shown in FIG. 3. Figure 2

[0055] Step S2-2, import environmental information, initialize sampling space, start pose, goal pose, gravitational coefficient, repulsive coefficient, and sampling step length, and the like, and establish an artificial potential field.

[0056] After importing the target region environmental information in the two-dimensional grid graph, a safety distance inflation process is performed on the obstacles. That is, according to the unmanned aerial vehicle body size, flight attitude error, and environmental safety redundancy requirement, a preset safety threshold distance is expanded outward on the basis of the original contour of the obstacle to form an equivalent model of the obstacle including a safety buffer area. Thus, it is ensured that even if the flight path planned subsequently deviates slightly due to algorithm calculation accuracy or actual flight disturbance, a safety gap can be maintained with the obstacle entity to avoid collision risk.

[0057] ​​After the safe inflation of the obstacle, the start_pose and goal_pose coordinates of the path planning in the grid map are defined; at the same time, the key parameters of the artificial potential field and the random tree growth are configured, including the gravity gain coefficient (regulating the attraction strength of the target point to the sampling point), the repulsion coefficient (adjusting the repulsion strength of the obstacle to the sampling point), the sampling step (controlling the single distance of the random tree node expansion), and the target bias rate (dynamically adjusting the probability of tilting towards the target point in the random sampling process), and the artificial potential field is established based on the above parameters.

[0058] The expression of the artificial gravitational potential field in space is:

[0059]

[0060] wherein, represents the gravitational function, represents the gravity gain coefficient, represents the position of the current node of the random tree expansion in space, is the position of the target point in space. represents the Euclidean distance from the current point in space to the target point.

[0061] The size of the gravitational force can be represented as the negative gradient of the gravitational field function, that is, the gravitational function can be expressed as:

[0062]

[0063] wherein, is the gravitational function, is the gradient operator, which is used to solve the direction of the fastest change of the potential field (i.e. the direction of the force).

[0064] The expression of the artificial repulsive force field in space is:

[0065]

[0066] wherein, represents the repulsive function, represents the repulsive gain coefficient, is the influence range of the obstacle, if it exceeds the range, the repulsive force field is 0.

[0067] The size of the repulsive force can be represented as the negative gradient of the repulsive field function, that is, the repulsive function can be expressed as:

[0068]

[0069] wherein, is the repulsive function.

[0070] The expression of the attraction of the random sampling point in space to the position of the current node of the random tree expansion in space is:

[0071]

[0072] wherein, the bias control of the random sampling point, is the random sampling point, is the position of the current node in space in the random tree expansion. Its force size can be expressed as This component enhances the robustness and path search efficiency of the algorithm in complex environments by introducing controllable randomness.

[0073] By presetting the weight proportion of the attractive force and the repulsive force in the combined field through the attractive force coefficient and the repulsive force coefficient, the expansion direction of the current node is effectively guided - both the attractive force component drives the sampling point to approach the target point, and the repulsive force component restricts the sampling point from the obstacle region.

[0074] The coupling force of the simulated wind field and the artificial potential field is the resultant force of the attractive force of the target point and the sampling point, the repulsive force of the obstacle and the wind field force, and the expression is:

[0075]

[0076] wherein, is the coupling force, is the attractive force of the random sampling point, is the attractive force of the target point to the current node, is the repulsive force of the obstacle to the current node, is the wind field force, is the fixed wind field intensity coefficient (used to adjust the influence weight of the wind field force in the comprehensive coupling effect).

[0077] Step S3, starting from the starting point, randomly sampling in the sampling space, and controlling the growth direction of the next expansion of the random tree by the coupling force of the simulated wind field and the artificial potential field, and generating a new node along the direction.

[0078] An adaptive bias sampling strategy is adopted, which dynamically adjusts the bias probability according to the distance from the current node of the random tree expansion to the target point, and the expression is:

[0079]

[0080] wherein, is the target bias probability, is the minimum target bias probability, is the maximum target bias probability, is the position of the starting point in space, is the initial distance from the starting point to the target point.

[0081] In the initial stage of path search, the distance to the target point is calculated using the starting point as a reference point. At this stage, the distance between the two points is relatively large, so the probability of bias towards the target point is low, and the algorithm tends to randomly sample and explore the environment. As sampling progresses, the distance from the newly expanded nodes to the target point gradually shortens, and the algorithm gradually increases the sampling probability biased towards the target point. This makes it more inclined to use the direction of the target point as the direction of the next sampling, thereby achieving a faster convergence speed.

[0082] like Figure 3 As shown, random sampling points are determined based on the target bias probability P. Then, through calculation, the resultant force that can affect the repulsive force of the obstacle, the target point, and the random sampling points are determined. The attractive force and wind force are used to calculate the resultant force of the four forces using a formula. The direction of this resultant force is the final growth direction. New nodes are obtained by extending a fixed step length in the growth direction. .

[0083] Specifically, according to the triangle rule for vector addition and subtraction, align the endpoint of the first vector with the starting point of the second vector. Point from the starting point of the first vector to the endpoint of the second vector; the resulting vector is the sum vector. Calculate the resultant force of all repulsive forces. Next, calculate the resultant force of the gravitational forces. Finally, superimpose the four forces according to the above rules to obtain the final resultant force and its direction.

[0084] Step S4: Perform collision detection on the new node to determine the optimal new node, and reconnect based on the optimal new node;

[0085] Perform collision checks on the generated line segments, i.e., determine... and Check if the line segment and the obstacle intersect. If they intersect, discard the newly sampled node. Return to step S3 and repeat. The sampling continues until no collisions are found, at which point the optimal new node is determined, and the new node is... Add to the random tree for the next sampling. .

[0086] After obtaining the optimal new node Then, within a radius of twice the step size centered at that node, the node cost is compared. For each adjacent node, the cost of the new node is calculated when that node is the parent node. The total path length to the starting point is calculated, and the node with the shortest path length is selected as the optimal parent node. Finally, Connect to the optimal parent node and update the path.

[0087] Repeat step S3 until the area around the target point is sampled.

[0088] Step S5, determining a path from the target point to the last sampling point around the target point Backtracking from the starting point to determine the path, and pruning and smoothing to complete the path planning.

[0089] A feasible path is obtained by backtracking from the parent node to the starting point.

[0090] As shown in Figure 4 , the optimized path can implement pruning processing on the feasible path.

[0091] Specifically, starting from the starting point, the current node i is checked along the generated path, and from the second node after it, the node j is selected and the line segment between node i and node j is constructed; the line segment is detected for obstacle collision to determine whether the line segment intersects with any obstacle, and if the line segment does not intersect with all obstacles, i.e. no collision occurs, it is indicated that the intermediate nodes (i+1 to j-1) between node i and node j are redundant nodes, and the redundant nodes are deleted from the feasible path.

[0092] As shown in Figure 5 , the path generated by the above process as shown in Figure 5 (a) is smoothed, the smoothing radius is calculated according to the length of the adjacent line segments, the transition curve is generated at the path inflection point through a 3rd order Bezier curve, so that the path is smoother, as shown in Figure 5 (b), thereby obtaining a flight route suitable for execution by the unmanned aerial vehicle.

[0093] Specifically, for each intermediate vertex of the pruned path, the lengths of the adjacent front and rear line segments are calculated, and the smaller value of the two line segment lengths and the product of the preset radius factor (the value range is 0.1~0.5) are taken as the smoothing radius at the vertex, to ensure the rationality and safety of the smooth transition.

[0094] The unit direction vector of each intermediate vertex line segment is calculated, and based on the vector and the above smoothing radius, the transition starting point and the transition ending point of the Bezier curve at the vertex are determined: the transition starting point is the position of the current vertex after moving the smoothing radius in the reverse direction along the unit vector of the front line segment, and the transition ending point is the position of the current vertex after moving the smoothing radius in the forward direction along the unit vector of the rear line segment.

[0095] The transition starting point and the transition ending point are used to generate two control points of the Bezier curve: the first control point is located at the position of the transition starting point moving in the forward direction along the unit vector of the front line segment ( ×0.33), and the second control point is located at the position of the transition ending point moving in the reverse direction along the unit vector of the rear line segment ( ×0.33).

[0096] By the Bezier curve formula, 11 interpolation points are generated between the transition start point and the end point of each vertex, and all interpolation points are sequentially connected with the start point and the end point of the path to form a continuous and smooth optimized path, as shown in Figure 6 .

[0097] Specifically, the calculation process is as follows:

[0098] Let the initial path after pruning be composed of a series of discrete vertices in order, and the vertex set is represented as , wherein is the start point of the path, is the end point of the path, is the i-th intermediate vertex of the path , and the coordinates of the i-th intermediate vertex are . The length of the forward line segment is , and the length of the backward line segment is . Based on the lengths of the line segments, the smoothing radius at the intermediate vertex is , .

[0099] The unit vector of the forward line segment is , and the unit vector of the backward line segment is .

[0100] The Bezier curve parameter is uniformly divided into 10 equal parts to obtain 11 discrete parameter values , wherein . Each parameter value is brought into the Bezier curve equation to obtain the corresponding 11 interpolation points , wherein .

[0101] Embodiment Two

[0102] In one or more embodiments, a UAV path planning system coupling a wind field and a potential field method is disclosed, specifically comprising:

[0103] A data acquisition module configured to acquire air pressure data and environmental data of a target area;

[0104] A grid construction module configured to construct a two-dimensional grid map of the target area, synchronously import the air pressure data and the environmental data, calculate the gradient of the grid point air pressure based on the air pressure data to obtain a simulated wind field, and simultaneously initialize a sampling space based on the environmental data and establish a potential field;

[0105] ​The node growth module is configured to: start from the starting point, randomly sample in the sampling space, control the growth direction of the next expansion of the random tree based on the coupling force of the simulated wind field and the artificial potential field, and generate a new node along the direction;

[0106] The node determination module is configured to: perform collision checking on the generated new node, if there is no collision, determine an optimal parent node, and perform reconnection; if collision occurs, re-sample until the target point around is sampled;

[0107] The path determination module is configured to: backtrack from the last sampling point around the target point to the starting point to determine a path, and perform pruning and smoothing to complete path planning.

[0108] Embodiment three

[0109] The embodiment provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and running on the processor, when the computer instructions are run by the processor, the steps of the UAV path planning method based on the coupling of the wind field and the artificial potential field are completed.

[0110] Embodiment four

[0111] The embodiment provides a computer readable storage medium for storing computer instructions, when the computer instructions are executed by the processor, the steps of the UAV path planning method based on the coupling of the wind field and the artificial potential field are completed.

[0112] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system) and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of the flows and / or blocks in the flowchart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that realizes the functions specified in the flowchart and / or block diagram. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The functions specified in one flow or multiple flows and / or blocks

[0113] These computer program instructions can also be stored in a computer readable storage medium, which can guide the computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer readable storage medium produce a product including instruction devices, which realize the functions specified in the flowchart and / or block diagram. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The functions specified in one flow or multiple flows and / or blocks

[0114] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0115] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0116] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for unmanned aerial vehicle path planning coupling a wind field with a potential field method, characterized in that, The application relates to a method for planning a path in a target area, and belongs to the field of path planning. The method comprises the following steps: acquiring air pressure data and environmental data of the target area; constructing a two-dimensional grid graph of the target area, synchronously importing the air pressure data and the environmental data, calculating the gradient of air pressure of each grid point based on the air pressure data to obtain a simulated wind field, importing the air pressure data, calculating the gradient vector of air pressure of each grid point in the two-dimensional plane to generate an air pressure gradient graph of the region; calculating the gradient modulus and normalizing the gradient vector to obtain a normalized gradient vector; applying the normalization method to each point in the region grid according to the grid index to generate a normalized gradient matrix of the whole region, extracting corresponding gradient components from the normalized gradient matrix, taking the negative of the extracted normalized gradient vector components to obtain the direction feature of the wind field at the grid point, and taking the modulus of the negative vector as the normalized representation of the wind field intensity, so as to determine the wind field force vector of each grid point, realize quantitative description of the wind field distribution of the target area, initialize a sampling space based on the environmental data, and establish an artificial potential field; starting from a starting point, randomly sampling in the sampling space, controlling the growth direction of the next expansion of a random tree based on the coupling force of the simulated wind field and the artificial potential field, and generating a new node along the direction; carrying out collision detection on the generated new node, determining an optimal parent node if there is no collision, and carrying out reconnection; if collision occurs, re-sampling until the target point around is sampled; backtracking from the last sampling point around the target point to the starting point to determine a path, and carrying out pruning and smoothing to complete path planning; In the formula, is the coupling force, is the gravity of the random sampling point, is the gravity of the target point on the current node, is the repulsion of the obstacle on the current node, is the wind field force, is the fixed wind field intensity coefficient; calculating the coupling force of the simulated wind field and the artificial potential field by calculating the resultant force of the attractive force of the target point and the sampling point on the current expansion node, the repulsive force of the obstacle and the wind field force, and the coupling force is a growth direction; the specific generation process of the new node comprises the following steps: an adaptive bias sampling strategy is adopted, the target bias probability is dynamically adjusted according to the distance from the current node of the random tree expansion to the target point, and a random sampling point is determined according to the target bias probability; the coupling force of the simulated wind field and the artificial potential field is obtained by calculating the resultant force of the attractive force of the target point and the sampling point on the current expansion node, the repulsive force of the obstacle and the wind field force, and the coupling force is a growth direction; a new node is obtained by extending a fixed step length according to the growth direction; wherein is the target bias probability, is the minimum target bias probability, is the maximum target bias probability, is the position of the start point in space, is the position of the target point in space, is the initial distance of the start point to the target point, denotes the position of the current node of the random tree expansion in space. 2.The method of claim 1, wherein, the expression of the target bias probability is as follows: 3.The method of claim 1, wherein, the establishment of the artificial potential field comprises the following steps: environmental information is imported in the two-dimensional grid graph, the safety distance of the obstacle is expanded, the sampling space, the starting point, the target point, the attractive force coefficient, the repulsive force coefficient and the sampling step are initialized, the attractive force field of the target point and the random sampling point and the repulsive force field of the obstacle are established, and the influence range of the repulsive force field is defined. the pruning and smoothing complete path planning, which comprises the following steps: pruning processing is carried out on the feasible path to optimize the path, starting from the starting point, judging along the generated path, checking the current node i, selecting nodes j from the second node of the current node i, and constructing a line segment between the node i and the node j; the line segment is subjected to obstacle collision detection, and it is judged whether the line segment intersects with any obstacle; if the line segment does not intersect with all obstacles, no collision occurs, and the intermediate node between the node i and the node j is a redundant node, and the redundant node is deleted. Then the path is smoothed, the smoothing radius is calculated according to the length of adjacent line segments, and a transition curve is generated at the path inflection point through a cubic Bezier curve.

4. An unmanned aerial vehicle path planning system coupling a wind field with a potential field method, adopting the unmanned aerial vehicle path planning method coupling a wind field with a potential field method according to any one of claims 1-3, characterized in that, The method comprises the steps that: The data acquisition module is configured to acquire air pressure data and environmental data of a target area; The grid construction module is configured to construct a two-dimensional grid map of the target area, synchronously import the air pressure data and the environmental data, calculate the gradient of the grid point air pressure based on the air pressure data to obtain a simulated wind field, and simultaneously initialize a sampling space and establish an artificial potential field based on the environmental data; The node growth module is configured to start from a starting point, perform random sampling in the sampling space, control the growth direction of the next extension of the random tree based on the coupling force of the simulated wind field and the artificial potential field, and generate a new node along the direction; The node determination module is configured to perform collision checking on the generated new node, determine an optimal parent node if there is no collision, and perform reconnection; if collision occurs, resample until the target point around is sampled; The path determination module is configured to backtrack from the last sampling point around the target point to the starting point to determine a path, and perform pruning and smoothing to complete path planning.

5. An electronic device, comprising: The computer instructions are executed by the processor to complete the UAV path planning method of the coupling wind field and artificial potential field method according to any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer instructions are executed by the processor to complete the UAV path planning method of the coupling wind field and artificial potential field method according to any one of claims 1-3.

Images