Path planning method and system based on random sampling of equipotential lines

Abstract

The present invention provides a path planning method based on random sampling of equipotential lines, constructs a sampling potential field function according to the position of obstacles, constructs a map potential field through the sampling potential field function, sets an equipotential line distribution coefficient k in the sampling potential field function, and adjusts the equipotential line distribution coefficient k to achieve the purpose of adjusting the node distribution; then, node sampling is performed on the equipotential lines, and a node optimization method based on a repulsive field is used, and finally, a graph search algorithm is used to perform an optimal path query on the constructed route map R, and a better path between the starting point and the end point is planned. The present invention improves the node sampling part of the PRM algorithm, and proposes a node sampling strategy based on equipotential line sampling and a node position optimization method based on a repulsive field, which can effectively increase the number of node sampling in narrow channels, optimize the quality of node sampling, and improve the success rate and quality of path planning.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile robot path planning, and more specifically to a mobile robot path planning method, device and equipment based on equipotential line sampling. Background Art

[0002] In the process of robots performing specific tasks, path planning is the core link, which determines the efficiency and safety of robot operation execution. Reasonable path planning is a crucial part of the task.

[0003] At present, there are many types of mobile robot path planning algorithms, including graph search-based, random sampling-based, artificial potential field method and bionic intelligence-based methods. These algorithms have been widely used in various scenarios and have achieved good performance. However, when faced with complex maps with narrow passages such as substations, traditional path planning algorithms often face many difficulties and challenges. For example, although the graph search-based algorithm can search for paths in narrow passages, the search speed becomes very slow when the map size is too large. The path search algorithm based on random sampling has a high search efficiency, but because its sampling method is random sampling in free space, it often cannot search points in narrow passages, which may result in too high a path cost. Summary of the invention

[0004] The technical problem to be solved by the present invention is to improve the node sampling part of the PRM algorithm, and propose a node sampling strategy based on equipotential line sampling and a node position optimization method based on repulsive field, which can effectively increase the number of node sampling in narrow channels and optimize the quality of node sampling, thereby improving the path planning success rate and path planning quality of the algorithm.

[0005] The present invention solves the above technical problems through the following technical means:

[0006] The path planning method based on random sampling of equipotential lines includes the following steps:

[0007] A sampling potential field function is constructed according to the position of the obstacle, and a map potential field is constructed through the sampling potential field function; the constructed map potential field shows a trend that the potential field gradient decreases faster around the obstacle and the gradient decreases slower at other positions far away from the obstacle. The equipotential line distribution coefficient k is set in the sampling potential field function, and the gradient decrease speed of the sampling potential energy is related to the sampling potential field distribution coefficient k, so that the distribution density of the equipotential lines and the node sampling probability are adjusted based on the sampling potential field distribution coefficient k; then, path node sampling is performed on the equipotential lines to obtain sampling points, and the positions of the sampling points are optimized based on the node optimization method of the repulsive field. Finally, edges are formed by connecting the optimized sampling points, and a path graph R is formed together with these sampling points; then, a search is performed using a graph search algorithm to find a reachable path between the starting point and the end point;

[0008] The specific process of constructing the sampling potential field and equipotential lines is as follows: Construct the sampling potential field function according to the obstacle position:

[0009]

[0010] Among them, U(X) represents the sampled potential energy at position X; Xobs is the position of the obstacle, and function d represents the distance between position X and obstacle Xobs; k is the equipotential line distribution coefficient, which determines the density of equipotential lines around the obstacle; the larger the k value, the denser the equipotential lines around the obstacle;

[0011] Path point sampling and optimization details:

[0012] 1) Random sampling of nodes on equipotential lines

[0013] Node sampling is performed on equipotential lines, and the sampling points are defined as:

[0014] Pnodes=s(L,np)

[0015] Where np is the number of sampling points; function s represents the collection of np sampling points on the equipotential line L, and the node sampling is completed on the equipotential line using the random sampling method;

[0016] 2) Optimization of sampling point location

[0017] Considering the obstacle as the repulsive force source, the sampling point is moved under the action of the repulsive force field; the moving distance of the sampling point to be optimized is expressed as:

[0018]

[0019] Among them, the function represents the moving distance of the sampling point P, which is a random number between D0 and Dmax; D0 is the minimum moving distance of the sampling point, which is set according to actual needs; the maximum distance that the sampling point moves under the influence of the repulsive field is defined as:

[0020]

[0021] in, r is the repulsive field gain coefficient, p is the maximum distance affected by obstacles; when the distance between the sampling point and the obstacle is greater than p , no movement will occur.

[0022] Furthermore, the graph search algorithm is an A* algorithm.

[0023] Furthermore, the potential energy change between any two adjacent equipotential lines is the same; the potential energy change is expressed as:

[0024] E = (Umax − Umin) / nl

[0025] The equipotential lines are defined as:

[0026] L=f(Umax,Umin,nl)

[0027] Among them, Umax and Umin are the highest and lowest sampling potential energies in the map respectively, L stores all points on the equipotential lines, which will be used for node sampling, and nl is the number of equipotential lines. The larger the k value, the denser the equipotential lines around the obstacle, and the greater the probability of being sampled.

[0028] The present invention also provides a path planning system based on random sampling of equipotential lines, comprising:

[0029] The sampling potential field and equipotential line construction module is used to construct a sampling potential field function according to the obstacle position, and to construct a map potential field through the sampling potential field function; the constructed map potential field shows a trend that the potential field gradient decreases faster around the obstacle and the gradient decreases slower at other locations far away from the obstacle. The equipotential line distribution coefficient k is set in the sampling potential field function. The gradient decrease speed of the sampling potential energy is related to the sampling potential field distribution coefficient k, so that the distribution density of the equipotential lines and the node sampling probability are adjusted based on the sampling potential field distribution coefficient k;

[0030] The path point sampling and optimization module performs path node sampling on equipotential lines to obtain sampling points, and optimizes the positions of sampling points based on the node optimization method of the repulsive field;

[0031] The path graph construction and path search module forms edges by connecting the optimized sampling points, and together with these sampling points, forms a path graph R; then it searches for a reachable path between the starting point and the end point by using a graph search algorithm;

[0032] The specific process of constructing the sampling potential field and equipotential lines is as follows: Construct the sampling potential field function according to the obstacle position:

[0033]

[0034] Among them, U(X) represents the sampled potential energy at position X; Xobs is the position of the obstacle, and function d represents the distance between position X and obstacle Xobs; k is the equipotential line distribution coefficient, which determines the density of equipotential lines around the obstacle; the larger the k value, the denser the equipotential lines around the obstacle;

[0035] Path point sampling and optimization details:

[0036] 1) Random sampling of nodes on equipotential lines

[0037] Node sampling is performed on equipotential lines, and the sampling points are defined as:

[0038] Pnodes=s(L,np)

[0039] Where np is the number of sampling points; function s represents the collection of np sampling points on the equipotential line L, and the node sampling is completed on the equipotential line using the random sampling method;

[0040] 2) Optimization of sampling point location

[0041] Considering the obstacle as the repulsive force source, the sampling point is moved under the action of the repulsive force field; the moving distance of the sampling point to be optimized is expressed as:

[0042]

[0043] Among them, the function represents the moving distance of the sampling point P, which is a random number between D0 and Dmax; D0 is the minimum moving distance of the sampling point, which is set according to actual needs; the maximum distance that the sampling point moves under the influence of the repulsive field is defined as:

[0044]

[0045] in, r is the repulsive field gain coefficient, p is the maximum distance affected by obstacles; when the distance between the sampling point and the obstacle is greater than p , no movement will occur.

[0046] Furthermore, the path graph construction and path search process are specifically as follows: the graph search algorithm is the A* algorithm.

[0047] Furthermore, the potential energy change between any two adjacent equipotential lines is the same; the potential energy change is expressed as:

[0048] E = (Umax − Umin) / nl

[0049] The equipotential lines are defined as:

[0050] L=f(Umax,Umin,nl)

[0051] Among them, Umax and Umin are the highest and lowest sampling potential energies in the map respectively, L stores all points on the equipotential lines, which will be used for node sampling, and nl is the number of equipotential lines. The larger the k value, the denser the equipotential lines around the obstacle, and the greater the probability of being sampled.

[0052] The advantages of the present invention are:

[0053] First, the present invention proposes a sampling potential field function, and the constructed map potential field shows a trend that the potential field gradient decreases faster around obstacles and the gradient decreases slower at other locations far away from obstacles. Moreover, we set the equipotential line distribution coefficient k in the sampling potential field function, and the purpose of adjusting the node distribution can be achieved by adjusting the equipotential line distribution. Under the influence of the sampling potential energy, the probability of being sampled around obstacles is high, and the probability of nodes being sampled in narrow channels is the highest. Then, a node optimization method based on a repulsive field is proposed, in which nodes move under the action of repulsion, and the repulsion they receive is related to the distance between them and the obstacles. Nodes that are close to obstacles will move farther, so that the nodes in the free space can be discretized and distributed more evenly. Finally, a graph search algorithm is used to query the optimal path of the constructed route map R, and efficiently plan the optimal path between the starting point and the end point.

[0054] The uniqueness of the algorithm described in the present invention is the construction of a sampling potential field function and a node optimization method based on a repulsive field. In the node selection stage, the probability of node sampling around obstacles can be adaptively increased, especially nodes in narrow spaces are most likely to be collected. The repulsive field function will affect the sampled nodes, further optimize the node position, so that the nodes in the narrow channel are concentrated in the center of the narrow channel, and the nodes in other free spaces are more evenly distributed, thereby improving the quality of path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A path planning map in an embodiment of the present invention;

[0056] Figure 2 It is a sampling potential field diagram in an embodiment of the present invention;

[0057] Figure 3 equipotential lines under different equipotential line distribution coefficients in the embodiment of the present invention;

[0058] Figure 4 is the sampling node distribution under different equipotential line distribution coefficients in the embodiment of the present invention;

[0059] Figure 5 Node optimization results under different repulsive field gain coefficients in the embodiment of the present invention;

[0060] Figure 6 This is the EPL-PRM path planning process in the embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] The technical solution of this embodiment is described in detail below in conjunction with the accompanying drawings. Figure 1 As shown in the figure, there are four cases: ordinary obstacles, ordinary narrow passages, long curved narrow passages, and complex curved narrow passages. These maps are all 800×800 pixel grid maps, which are used to test the performance of the algorithm in different scenarios.

[0063] Step 1: Construct sampling potential field and equipotential lines.

[0064] In the node sampling stage, it is hoped that more nodes can be collected around obstacles or in narrow channels to improve the quality of path planning. The sampling potential field function constructed according to the obstacle position is:

[0065]

[0066] Among them, U(X) represents the sampled potential energy at position X. In the entire map, the obstacle area is the area with the maximum potential energy, and the sampled potential energy is 1. The minimum potential energy is at the farthest position from the obstacle, which is a value close to 0. Xobs is the position of the obstacle, and the function d represents the distance between position X and obstacle Xobs. k is the equipotential line distribution coefficient, which determines the density of equipotential lines around the obstacle. The larger the k value, the denser the equipotential lines around the obstacle.

[0067] The node sampling is performed on uniform equipotential lines, and the potential energy change between any two adjacent equipotential lines is the same. The potential energy change can be expressed as:

[0068] E = (Umax − Umin) / nl

[0069] The equipotential lines are defined as:

[0070] L=f(Umax,Umin,nl)

[0071] Among them, Umax and Umin are the highest and lowest sampling potentials in the map, respectively. L stores all points on the equipotential lines, which will be used to sample the probability map nodes. nl is the number of equipotential lines. During the equipotential line sampling process, the more equipotential lines there are, the more positions can be sampled, and the sampled nodes in the free space will be more evenly distributed. However, too many equipotential lines will increase the amount of calculation and reduce planning efficiency.

[0072] The sampling potential field function proposed in the present invention relates the potential energy drop speed of the current position to the distance between it and the obstacle. In the area close to the obstacle, there is a greater sampling potential energy and the potential energy drop is more drastic. This part of the free space has a higher probability of being sampled during the node sampling stage. Figure 2 As shown, the present invention uses color and height to represent the sampling potential field of the four maps. The obstacle has the largest sampling potential energy, and there are obstacles on both sides of the narrow channel, forming a potential energy canyon.

[0073] The sampling potential field function described in the present invention is used to establish a global potential map. Before constructing the equipotential lines of the sampled nodes, it is necessary to adjust the adjustable parameter k in the potential field construction function. The equipotential line construction function is an inverse proportional function. The larger the exponent k, the faster the potential field gradient decreases near the obstacle. Therefore, the exponent k can be used as the equipotential line distribution coefficient. By adjusting this parameter, the distribution of equipotential lines in the map can be changed, thereby affecting the distribution state of the nodes. Figure 3 As shown in the figure, the construction of equipotential lines under different k values ​​is shown, where the color of the equipotential line represents the potential energy of the nodes on the equipotential line (the sampling potential energy varies from 0 to 1). In this experiment, the values ​​of k are set to 0.1, 0.5 and 1 respectively, and the number of equipotential lines nl is 40.

[0074] Step 2: Random sampling of nodes on equipotential lines.

[0075] In the sampling stage, this algorithm retains the random sampling characteristics of the PRM algorithm, which is conducive to improving the execution speed of the algorithm and has probabilistic completeness. As the number of sampling points increases, the path planning success rate will approach 1 infinitely. Node sampling is performed on the equipotential line, and the sampled nodes can be defined as:

[0076] Pnodes=s(L,np)

[0077] Wherein, np is the number of sampling points. Function s represents the collection of np number of path points on the equipotential line L, and the node sampling is completed on the equipotential line using a random sampling method. Generally speaking, the more path points are collected, the greater the probability that the robot finds the optimal path, but the constructed roadmap is also more complex, and the path search speed may also decrease accordingly. The present invention constructs a sampling potential energy function and extracts equipotential lines based on potential energy changes. Random sampling on equipotential lines can make the sampling points around obstacles more dense.

[0078] Nodes are randomly sampled on equipotential lines. k determines the distribution of equipotential lines, which in turn affects the sampling of nodes. Figure 4 The node sampling results of the four maps under different k values ​​are shown in Figure 1. The nodes around obstacles in the map are relatively dense. The nodes in the narrow passage are the most dense per unit area. As k increases, there are more nodes around obstacles, which also leads to a decrease in nodes in other free spaces. When k is equal to 1, most nodes are concentrated around obstacles, which increases the number of nodes distributed in narrow passages and improves the success rate of path planning.

[0079] exist Figure 4 As shown in (a), in a simple obstacle map, a larger k value causes nodes to be concentrated around obstacles, which is not conducive to path planning. In this map, it is more reasonable to choose a smaller k value. In a complex narrow channel environment, it is difficult for ordinary samplers to collect nodes in the narrow channel, resulting in path planning failure or failure to find the optimal path. Therefore, in a narrow channel map environment, the k value can be increased according to the complexity of the narrow channel. For example, in Figure 4 Middle (c), Figure 4 In (d), or even more complex cases, a larger k value may be considered to increase the number of waypoint sampling in the narrow channel.

[0080] Step three: node location optimization.

[0081] The present invention proposes an optimization method based on artificial potential field, which regards obstacles as repulsive sources and makes sampling points move under the action of repulsive fields. In order to maintain the probability completeness of the algorithm, randomness is still retained, and the maximum and minimum moving distances of sampling points are limited. The moving direction of the sampling node is always along the gradient descent direction of the repulsive field, which makes the nodes in the narrow channel attracted to the central area, thereby greatly improving the quality of path planning.

[0082] It is hoped that the nodes closer to the obstacles will move farther after optimization. The distance moved by the node to be optimized can be expressed as:

[0083]

[0084] Among them, the function Indicates the moving distance of node P, which is a random number between D0 and Dmax. D0 is the minimum moving distance of the node, which can be set according to actual needs and can be set to 0 in general. The maximum distance that a node can move under the influence of the repulsive field can be defined as:

[0085]

[0086] in, r is the repulsive field gain coefficient, p is the maximum distance affected by obstacles. p , no movement will occur.

[0087] The repulsive force on the node will gradually change during the movement, but the moving distance is only affected by the repulsive force at its initial position and only needs to be calculated once based on its initial position. In path planning, the node will move along the gradient direction under the influence of the repulsive field to achieve the purpose of staying away from obstacles.

[0088] like Figure 5 , showing the different repulsion field gain coefficients in two map environments r The result of node optimization. The randomness of the PRM algorithm is retained. The distance that the node moves under the repulsive field is random. The maximum distance of random movement is related to the repulsive force and repulsive gain coefficient. r The direction of movement is always along the descending direction of the repulsive potential energy gradient. The original sampling node data is r = 0. As the repulsion field coefficient increases, the nodes are more evenly distributed in the free space, and the nodes in the narrow channel are concentrated in the central area. When choosing the repulsion field gain coefficient, you can choose it according to the complexity of the map. r The larger the value, the farther the obstacle can move under the influence of the repulsion field. More complex narrow channel maps often require a larger repulsion gain coefficient.

[0089] Step 4: Path map construction and path search.

[0090] The route map and path search method constructed by the present invention are similar to the PRM algorithm, that is, by connecting the barrier-free nodes to form edges, and together with these nodes, a path map is formed. Then, the A* algorithm is used to perform a graph search to find a reachable path between the starting point and the end point. Among the randomly sampled nodes, when the distance between two nodes is short enough and there are no obstacles between them, the two nodes are connected to form an edge.

[0091] Random sampling on equipotential lines can adjust the occupancy ratio of sampling points around obstacles and in free space, so that the planner can collect more critical path nodes. In addition, the node optimization strategy proposed in the present invention can further optimize the collected nodes, making their distribution in free space more even, and making the node distribution in narrow channels more concentrated in the center of the channel. At the same time, the present invention also retains some of the probabilistic characteristics of the planner, increasing the adaptability of the algorithm in different environments. The path planning algorithm based on random sampling of equipotential lines includes four steps. The first is the construction of equipotential lines and random path point sampling, followed by the optimization of path points under a repulsive field, followed by the construction of a probabilistic roadmap, and finally the search for the optimal path. The search effect of the algorithm was tested on four maps. The running process and results of the EPL-PRM algorithm are shown in the following figure. Figure 6 shown.

[0092] In maps with narrow channels, it is necessary to increase the number of sampling nodes as appropriate to increase the success rate of planning. At the same time, increase the equipotential line distribution coefficient to concentrate more points in narrow channels. In a normal obstacle environment, a smaller k value can be selected to make the node distribution more uniform. In this experiment, k is set to 0.2. In a scene with a narrow channel, a larger k value is selected to increase the probability of node sampling in the narrow channel. In this implementation, the k value of the narrow channel map is 0.5.

[0093] Sampling is one of the core steps of PRM algorithms. Currently, most improvements on PRM algorithms focus on the sampling link, so as to improve the quality of the constructed probabilistic roadmap, thereby improving the success rate, planning speed and planning quality of path planning. This embodiment proposes a path planning strategy EPL-PRM based on equipotential line sampling for the sampling link of PRM. This section will summarize its algorithm framework and give the pseudo code of the algorithm.

[0094]

[0095] First, this embodiment proposes a sampling potential field function, and the constructed map potential field shows a trend that the potential field gradient around the obstacle is faster, and the gradient of other locations far away from the obstacle is slower. In addition, we set the equipotential line distribution coefficient k in the sampling potential field function, and the equipotential line distribution can be adjusted to achieve the purpose of adjusting the node distribution. Under the influence of the sampling potential energy, the probability of being sampled around the obstacle is high, and the probability of nodes being sampled in narrow channels is the highest. Then, a node optimization method based on a repulsive field is proposed, in which the node moves under the action of repulsion, and the repulsion it receives is related to the distance between it and the obstacle. Nodes close to the obstacle will move farther, so that the nodes in the free space can be discretized and distributed more evenly. Finally, a graph search algorithm is used to query the optimal path of the constructed route map R, and efficiently plan the optimal path between the starting point and the end point.

[0096] The uniqueness of the algorithm described in this embodiment is the construction of a sampling potential field function and a node optimization method based on a repulsive field. In the node selection stage, the probability of node sampling around obstacles can be adaptively increased, especially nodes in narrow spaces are most likely to be collected. The repulsive field function will affect the sampled nodes, further optimize the node position, concentrate the nodes in the narrow channel in the center of the narrow channel, and distribute the nodes in other free spaces more evenly, so as to improve the quality of path planning.

[0097] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. 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 may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A path planning method based on random sampling of equipotential lines, characterized in that: The following steps are involved: A sampling potential field function is constructed according to the obstacle position, and a map potential field is constructed through the sampling potential field function; The constructed map potential field shows a trend that the potential field gradient decreases faster around obstacles and slower at other locations far away from obstacles. The equipotential line distribution coefficient k is set in the sampling potential field function. The gradient decrease speed of the sampling potential is related to the sampling potential field distribution coefficient k, so that the distribution density of the equipotential lines and the node sampling probability are adjusted based on the sampling potential field distribution coefficient k; then, path node sampling is performed on the equipotential lines to obtain sampling points, and the positions of the sampling points are optimized based on the node optimization method of the repulsive field. Finally, edges are formed by connecting the optimized sampling points, and the path graph R is formed together with these sampling points; then, a graph search algorithm is used to search for a reachable path between the starting point and the end point; The specific process of constructing the sampling potential field and equipotential lines is as follows: Construct the sampling potential field function according to the obstacle position: Among them, U(X) represents the sampled potential energy at position X; Xobs is the position of the obstacle, and function d represents the distance between position X and obstacle Xobs; k is the equipotential line distribution coefficient, which determines the density of equipotential lines around the obstacle; the larger the k value, the denser the equipotential lines around the obstacle; Path point sampling and optimization details: 1) Random sampling of nodes on equipotential lines Node sampling is performed on equipotential lines, and the sampling points are defined as: Pnodes=s(L,np) Where np is the number of sampling points; function s represents the collection of np sampling points on the equipotential line L, and the node sampling is completed on the equipotential line using the random sampling method; 2) Optimization of sampling point location Considering the obstacle as the repulsive force source, the sampling point is moved under the action of the repulsive force field; the moving distance of the sampling point to be optimized is expressed as: Among them, the function represents the moving distance of the sampling point P, which is a random number between D0 and Dmax; D0 is the minimum moving distance of the sampling point, which is set according to actual needs; the maximum distance that the sampling point moves under the influence of the repulsive field is defined as: in, r is the repulsive field gain coefficient, p is the maximum distance affected by obstacles; when the distance between the sampling point and the obstacle is greater than p , no movement will occur.

2. The path planning method based on random sampling of equipotential lines according to claim 1, characterized in that: The graph search algorithm is the A* algorithm.

3. The path planning method based on random sampling of equipotential lines according to claim 1, characterized in that: The potential energy change between any two adjacent equipotential lines is the same; the potential energy change is expressed as: E = (Umax − Umin) / nl The equipotential lines are defined as: L=f(Umax,Umin,nl) Among them, Umax and Umin are the highest and lowest sampling potential energies in the map respectively, L stores all points on the equipotential lines, which will be used for node sampling, and nl is the number of equipotential lines. The larger the k value, the denser the equipotential lines around the obstacle, and the greater the probability of being sampled.

4. A path planning system based on random sampling of equipotential lines, characterized in that: include: The sampling potential field and equipotential line construction module is used to construct the sampling potential field function according to the obstacle position, and to construct the map potential field through the sampling potential field function; The constructed map potential field shows a trend that the potential field gradient decreases faster around obstacles and decreases slower at other locations far away from obstacles. The equipotential line distribution coefficient k is set in the sampling potential field function. The gradient decrease speed of the sampling potential energy is related to the sampling potential field distribution coefficient k, so the distribution density of the equipotential lines and the node sampling probability are adjusted based on the sampling potential field distribution coefficient k. The path point sampling and optimization module performs path node sampling on equipotential lines to obtain sampling points, and optimizes the positions of sampling points based on the node optimization method of the repulsive field; The path graph construction and path search module forms edges by connecting the optimized sampling points [3], and together with these sampling points, forms a path graph R; then it searches by using a graph search algorithm to find a reachable path between the starting point and the end point; The specific process of constructing the sampling potential field and equipotential lines is as follows: Construct the sampling potential field function according to the obstacle position: Among them, U(X) represents the sampled potential energy at position X; Xobs is the position of the obstacle, and function d represents the distance between position X and obstacle Xobs; k is the equipotential line distribution coefficient, which determines the density of equipotential lines around the obstacle; the larger the k value, the denser the equipotential lines around the obstacle; Path point sampling and optimization details: 1) Random sampling of nodes on equipotential lines Node sampling is performed on equipotential lines, and the sampling points are defined as: Pnodes=s(L,np) Where np is the number of sampling points; function s represents the collection of np sampling points on the equipotential line L, and the node sampling is completed on the equipotential line using the random sampling method; 2) Optimization of sampling point location Considering the obstacle as the repulsive force source, the sampling point is moved under the action of the repulsive force field; the moving distance of the sampling point to be optimized is expressed as: Among them, the function represents the moving distance of the sampling point P, which is a random number between D0 and Dmax; D0 is the minimum moving distance of the sampling point, which is set according to actual needs; the maximum distance that the sampling point moves under the influence of the repulsive field is defined as: in, r is the repulsive field gain coefficient, p is the maximum distance affected by obstacles; when the distance between the sampling point and the obstacle is greater than p , no movement will occur.

5. The path planning system based on random sampling of equipotential lines according to claim 4, characterized in that: The path graph construction and path search process are specifically as follows: the graph search algorithm is the A* algorithm.

6. The path planning system based on random sampling of equipotential lines according to claim 4, characterized in that: The potential energy change between any two adjacent equipotential lines is the same; the potential energy change is expressed as: E = (Umax − Umin) / nl The equipotential lines are defined as: L=f(Umax,Umin,nl) Among them, Umax and Umin are the highest and lowest sampling potential energies in the map respectively, L stores all points on the equipotential lines, which will be used for node sampling, and nl is the number of equipotential lines. The larger the k value, the denser the equipotential lines around the obstacle, and the greater the probability of being sampled.

Images