A Robot Path Planning Method Based on an Improved RRT Algorithm

Abstract

This invention discloses a robot path planning method based on an improved RRT algorithm. This method limits the sampling area by using the maximum sampling radius and the newly added nodes from the previous time step, ensuring that the acquired sampling points are more likely to be biased towards the target point. Simultaneously, by restricting the sampling points, this invention can maximize the approach to the target point when there are few obstacles, and minimize the approach when there are many obstacles, reducing the number of sampling points when acquiring new nodes in the random tree and improving the efficiency of path generation. Furthermore, this invention removes redundant nodes from the path after generation and performs pruning operations on the nodes, optimizing the path nodes, reducing the number of corners in the generated path, and improving path planning performance. Compared to the RRTstar and Astar algorithms, this invention can effectively shorten the path length and improve search efficiency.

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, specifically relating to a robot path planning method based on an improved RRT algorithm. Background Technology

[0002] In recent years, mobile robots have been widely used in military operations, emergency rescue, warehouse management, and logistics transportation. Path planning is an indispensable and crucial component of mobile robots, typically referring to finding a collision-free and optimal path from a starting point to a destination. Currently, path planning algorithms can be broadly categorized into three types: search-based algorithms, bio-inspired algorithms, and sampling-based algorithms. Search-based path planning algorithms mainly include A*, Dijkstra's algorithm, and Theta* algorithm. These algorithms divide the map into a grid and then use a search algorithm to obtain the path from the starting point to the destination. Although these methods are computationally simple, the time it takes to find a path and the length of the path depend on the fineness of the grid division. Bio-inspired path planning algorithms mainly include Ant Colony Optimization (ACA), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO) algorithms. These algorithms have strong adaptability and parallelism, but their computation time is relatively long, and they are prone to getting trapped in local optima. Sampling-based path planning algorithms obtain a collision-free path from the starting point to the destination by randomly obtaining sampling points on a map and then connecting these points. These algorithms mainly include the Rapid Expanding Random Tree (RRT) algorithm, the Rapid Expanding Random Tree Star (RRT*) algorithm, and the Probabilistic Route Map (PRM) algorithm.

[0003] The RRT algorithm has too much randomness and blindness. The RRT* algorithm reduces path cost by finding parent nodes and rewiring, but increases search time and is prone to getting trapped in local optima. Summary of the Invention

[0004] In view of the excessive randomness and blindness of existing RRT algorithms, and the inability of RRT* algorithms to simultaneously solve the problems of path length and search time, the purpose of this invention is to provide an improved fast expanding random tree path planning algorithm, which is suitable for high-dimensional spaces and complex environments with non-convex obstacles.

[0005] The robot path planning method based on the improved RRT algorithm proposed in this invention includes the following steps:

[0006] Step 1: Obtain environmental information around the robot and construct a coordinate system. Environmental information includes the initial point Q. start Obstacles and target point Q goal Location information; initial point Q start Add it to the random tree as the root node.

[0007] Step 2: Obtain sampling point Q in the constructed coordinate system. sample Sampling point Q sample azimuth and sampling radius r limit The expression is:

[0008]

[0009] r limit =a3r max +a4r new +a5eps

[0010] Where a1 is the target guidance coefficient, a1>k max -k min ;θ initial Let Q be the initial point start and target point Q goal The angle formed by the line connecting the x-axis and the x-axis; k is a random number with an expected value of zero, k∈[k... min k max ];k min and k max These are the minimum and maximum values ​​of the preset random number k, respectively; a2 is the global coefficient in radians; a3 is the expansion coefficient; r max a4 is the maximum sampling radius; r is the global coefficient of the step size. new For the previous newly added node Q′ new With initial point Q start The Euclidean distance; a5 is a random step coefficient related to the sampling collision coefficient col, a5 = rand - col; rand is a random number within a preset range; col is the sampling collision coefficient; Q is a node on the random tree; esp is the maximum growth step size, its expression is:

[0011]

[0012] Where s is the preset minimum growth step size; x step This is the step size coefficient.

[0013] If the random tree only has a root node, then the root node is used as the previous newly added node Q′. new Based on sampling point Q sample Get random point Q rand .

[0014] Step 3: Traverse all nodes in the random tree and calculate the relationship between each node and the random point Q. rand Euclidean distance, obtain the distance to a random point Q. rand The nearest node is used as the feature node Q near Based on random point Q rand and feature node Q nearGet the new node Q new .

[0015] Step 4: If a new node Q is added new With feature node Q near If the connecting line collides with an obstacle, repeat steps two and three until a new node Q is obtained. new With feature node Q near The connection has no obstacles, and the newly acquired node Q will be obtained. new Add to the random tree. Update the step size coefficient x based on the number of collisions (num). step And the sampling collision coefficient col.

[0016] Step 5: Update the newly added node Q new parent node Q parent .

[0017] Step 6: Traverse the existing nodes in the random tree and calculate the distance between each node and the target point Q. goal The Euclidean distance between them; if there exists a node and the target point Q goal If the Euclidean distance between nodes is less than the minimum growth step size s, then that node is taken as the target point Q. goal The parent node, and from the target point Q goal Begin by selecting the parent node Q that connects each node in turn. parent Until the initial point Q start , as the initial point Q start Move to target point Q goal If the path is correct, the iteration process ends; otherwise, repeat steps two through five to continue expanding the random tree.

[0018] Preferably, in step two, if the robot can move in three-dimensional space, then sampling point Q is also constructed. sample The polar angle θ, the expression for the polar angle θ is:

[0019] θ=b1(θ initial +k)+2b2π

[0020] Where b1 is the target orientation coefficient, b1>k max -k min b2 is the global coefficient for radians.

[0021] Preferably, in step five, the initial point Q is obtained. start Move to target point Q goal After determining the path, remove redundant nodes from the path. The method for removing redundant nodes is as follows:

[0022] With target point Q goal As the endpoint, the initial point Q startAs the starting point, connect the ending point to the starting point; if the line connecting the starting and ending points collides with an obstacle, then the child node Q of the starting point is used as the starting point. child Update the starting point until the line connecting the starting and ending points does not collide with any obstacles. Use the updated starting point as the ending point, and the initial point is Q. start Starting from point Q, repeat the above process until the endpoint coincides with the initial point Q. start The line connecting the obstacles did not collide with them.

[0023] Preferably, in step five, after removing redundant nodes, each node in the path is selected sequentially as the mobile node Q. choose First move node Q choose Moving towards its parent node, the translation vector d1 for each move is:

[0024]

[0025] Until the moving node Q choose The connection between the node and its child nodes collides with an obstacle or moves node Q. choose After reaching its parent node, stop moving towards its parent node; then move the node Q. choose Move towards its child nodes, and the translation vector d2 for each move is:

[0026]

[0027] Until the moving node Q choose The connection between the node and its parent node collides with an obstacle or moves node Q. choose Once a node reaches its child node, stop moving towards its child nodes. After moving all nodes in the path, obtain the updated path as the initial point Q. start Move to target point Q goal The final path.

[0028] Preferably, in step two, if there is no potential field U during the sampling point acquisition process... req , with sampling point Q sample As a random point Q rand Conversely, random point Q rand The expression is:

[0029]

[0030] Among them, Q req F is the center point of the potential field; req For sampling point Q sample The repulsive force it receives.

[0031] Preferably, the sampling point Q sample The repulsive force F received req The expression is:

[0032]

[0033] Where, k req is the potential field gain coefficient; r is the radius of the potential field; n is the adjustment factor coefficient.

[0034] Preferably, the potential field U req The confirmation method is as follows:

[0035] If the neighbor set Q neighbor If the number of elements in the brackets exceeds a preset threshold, then add a new node Q. new Q, the center point of the potential field req The generation only affects the next random point Q. rand Potential field U req And reassign the maximum sampling radius r max and distance r new The maximum sampling radius r′ after assignment max and distance r′ new The expression is:

[0036] r′ max =r max -1.1r

[0037] r′ new =r new -1.1r.

[0038] Preferably, in step five, the newly added node Q is updated. new parent node Q parent The method is as follows:

[0039] Add a new node Q with the center of the circle. new All nodes within a circle with radius r equal to the radius of the potential field are considered as the neighbor set Q. neighbor []. Get the set of neighboring nodes Q neighbor From each node in [] to the newly added node Q new Total cost:

[0040] cost = cost neighbor +dist(Q neighbor -Q new )

[0041] Among them, Q neighbor Let Q be the set of neighboring nodes. neighbor The neighboring nodes in []; cost neighbor Let Q be the initial point start Reaching neighboring node Q neighbor The cost.

[0042] With the neighbor set Qneighbor From each node in [] to the newly added node Q new The node corresponding to the minimum cost is taken as the minimum cost node Q. min If the minimum cost node Q min With the newly added node Q new If there are no collisions, then the node with the minimum cost Q is selected. min As a newly added node Q new Parent node.

[0043] Preferably, in step three, the newly added node Q is obtained. new The method is as follows:

[0044]

[0045] Among them, dist(Q rand Q near ) represents a random point Q rand With feature node Q near The Euclidean distance between them.

[0046] Preferably, in step four, the step size coefficient x step The update method is as follows: in the original step size coefficient x step Based on this, add a small step size Δx step Infinite step size Δx step It decreases as the number of collisions num increases. The update method for the sampled collision coefficient col is as follows: a small coefficient Δcol is added to the original sampled collision coefficient col. The initial value of the small coefficient Δcol is 0. After two consecutive collisions, the small coefficient Δcol increases by a fixed value for each additional collision, until the small coefficient Δcol reaches the preset maximum value.

[0047] The beneficial effects of this invention are:

[0048] 1. This invention limits the sampling area by using the maximum sampling radius and the newly added nodes from the previous time step, so that the obtained sampling points can be biased towards the target point with a higher probability. At the same time, by limiting the sampling points, this invention can get as close to the target point as possible when there are few obstacles, and get as close to the target point as little as possible when there are many obstacles, thereby reducing the number of sampling points when obtaining new nodes in the random tree and improving the efficiency of path generation.

[0049] 2. This invention removes redundant nodes from the path after path generation and performs pruning operations on the nodes in the path, thereby optimizing the path nodes, reducing the number of turns in the generated path, and improving path planning performance. Compared with RRTstar and Astar algorithms, this invention can effectively shorten the path length and improve search efficiency. Attached Figure Description

[0050] Figure 1 This is the overall flowchart of the present invention.

[0051] Figure 2 This is a schematic diagram of the azimuth angle of a random point in this invention.

[0052] Figure 3 This is a schematic diagram of the sampling radius of random points in this invention.

[0053] Figure 4 This is a schematic diagram of random tree growth in this invention.

[0054] Figure 5 This is a schematic diagram of reselecting the parent node in this invention.

[0055] Figure 6 This is a schematic diagram of the escape potential field generated in this invention.

[0056] Figure 7 This is a schematic diagram illustrating the removal of redundant nodes in this invention.

[0057] Figure 8 This is a schematic diagram of pruning in this invention.

[0058] Figure 9 This is a simulation diagram of the present invention and the RRT* algorithm for path planning in complex maps.

[0059] Figure 10 This is a simulation diagram of the path planning of the present invention and the RRT* algorithm in a single-pass narrow road map. Detailed Implementation

[0060] The present invention will be further described below with reference to the accompanying drawings.

[0061] like Figure 1 As shown, a robot path planning method based on an improved RRT algorithm includes the following steps:

[0062] Step 1: Acquire environmental information around the robot and construct a coordinate system; if the robot moves in a two-dimensional plane, establish a planar coordinate system; if the robot moves in three-dimensional space, establish a three-dimensional Cartesian coordinate system. Environmental information includes the initial point Q. start Obstacles and target point Q goal Location information; initial point Q start Add it to the random tree as the root node.

[0063] Step Two: As Figure 2 , 3 As shown in Figure 4, the sampling point Q is obtained in the constructed coordinate system. sample If the robot moves in a two-dimensional plane, then the sampling point Q sample azimuth (the angle formed with the x-axis) and sampling radius r limit The expression is:

[0064]

[0065] r limit =a3r max +a4r new +a5eps

[0066] Where a1 is the target guidance coefficient, a1>k max -k min ;θ initial Let Q be the initial point start and target point Q goal The angle formed by the line connecting the x-axis and the x-axis; k is a random number with an expected value of zero, k∈[k... min k max ];k min and k max These are the minimum and maximum values ​​of the preset random number k, respectively; a2 is the global coefficient in radians; a3 is the expansion coefficient; r max r is the maximum sampling radius max =max(dist(Q,Q) start )); a4 is the global coefficient for step size; r new For the previous newly added node Q′ new With initial point Q start The Euclidean distance; a5 is a random step coefficient related to the sampling collision coefficient col, a5 = rand - col; rand is a random number within a preset range; col is the sampling collision coefficient; Q is a node on the random tree; max(·) is the maximum value; dist(Q,Q start ) represents the distance from a node in a random tree to the initial point Q. start The Euclidean distance; esp is the maximum growth step size, its expression is:

[0067]

[0068] Where s is the preset minimum growth step size; x step This is the step size coefficient, initially set to 0.

[0069] If the robot moves in three-dimensional space, then add sampling point Q to the two-dimensional space. sample The polar angle θ (the angle formed with the z-axis) is expressed as follows:

[0070] θ=b1(θ initial +k)+2b2π

[0071] Where b1 is the target orientation coefficient, b1>kmax -k min b2 is the global coefficient for radians.

[0072] In this embodiment, the target guidance coefficients a1 and b1 are 0.3; the minimum and maximum values ​​of the random number k are -0.1 and 0.1, respectively; the global coefficients a2 and b2 are 0.0583rand; and the random number rand takes values ​​in [0,3].

[0073] If the random tree only has a root node, then the root node is used as the previous newly added node Q′. new .

[0074] Step 3: Based on sampling point Q sample Get random point Q rand If there is no potential field U during the sampling point acquisition process req , with sampling point Q sample As a random point Q rand Conversely, random point Q rand The expression is:

[0075]

[0076] Among them, Q req F is the center point of the potential field; req For sampling point Q sample The repulsive force experienced is expressed as follows:

[0077]

[0078] Among them, ▽U req For the repulsive potential field U rep The gradient of k; req is the potential field gain coefficient, with a value of 5; r is the radius of the potential field, r = 3 × esp.

[0079] Potential field U req The expression is:

[0080]

[0081] Where n is the adjustment factor coefficient, and its value is 2.

[0082] Step Four: As Figure 4 As shown, traverse all nodes on the random tree and calculate the relationship between each node and the random point Q. rand Euclidean distance, obtain the distance to a random point Q. rand The nearest node is used as the feature node Q near Based on random point Q rand and feature node Q near Get the new node Q new Add a new node Qnew The expression is:

[0083]

[0084] Among them, dist(Q rand Q near ) represents a random point Q rand With feature node Q near The Euclidean distance between them.

[0085] Step 5: If a new node Q is added new With feature node Q near If the connecting line collides with an obstacle, repeat steps two and three until a new node Q is obtained. new With feature node Q near The connection has no obstacles, and the newly acquired node Q will be obtained. new Add to the random tree. To obtain the newly added node Q in the random tree. new During the process, the number of times random points are generated is used as the collision count num, and the step size coefficient x is updated based on the collision count num. step And the sampling collision coefficient col.

[0086] Step size coefficient x step The update method is as follows: in the original step size coefficient x step Based on this, add a small step size Δx step Infinite step size Δx step It decreases as the number of collisions num increases. The update method for the sampled collision coefficient col is as follows: add a small coefficient Δcol to the original sampled collision coefficient col. The initial value of the small coefficient Δcol is 0. After two consecutive collisions, Δcol is increased by 0.5 for each additional collision, until the small coefficient Δcol reaches the preset maximum value of 2.

[0087] Step Six: As Figure 5 As shown, with the center of the circle as the newly added node Q new All nodes within a circle with radius r equal to the radius of the potential field are considered as the neighbor set Q. neighbor []. Get the set of neighboring nodes Q neighbor From each node in [] to the newly added node Q new Total cost:

[0088] cost = cost neighbor +dist(Q neighbor -Q new )

[0089] Among them, Q neighbor Let Q be the set of neighboring nodes. neighbor The neighboring nodes in []; cost neighborLet Q be the initial point start Reaching neighboring node Q neighbor The cost.

[0090] With the neighbor set Q neighbor From each node in [] to the newly added node Q new The node corresponding to the minimum cost is taken as the minimum cost node Q. min If the minimum cost node Q min With the newly added node Q new If there are no collisions, then the node with the minimum cost Q is selected. min As a newly added node Q new Parent node. For example... Figure 5 As shown, if the total cost from point a to point c is less than the total cost from point b to point c, then point a is reselected as the parent node of point c.

[0091] Step Seven: As Figure 6 As shown, to escape a local optimum, if the set of neighboring nodes Q... neighbor If the number of elements in the brackets exceeds a preset threshold, then add a new node Q. new Q, the center point of the potential field req The generation only affects the next random point Q. rand Potential field U req And reassign the maximum sampling radius r max and distance r new The maximum sampling radius r′ after assignment max and distance r′ new The expression is:

[0092] r′ max =r max -1.1r

[0093] r′ new =r new -1.1r

[0094] The maximum sampling radius r′ after assignment max and distance r′ new The maximum sampling radius r of the next sampling node max and distance r new .

[0095] Step 8: Traverse the existing nodes in the random tree and calculate the distance between each node and the target point Q. goal The Euclidean distance between them; if there exists a node and the target point Q goal If the Euclidean distance between nodes is less than the minimum growth step size s, then that node is taken as the target point Q. goal The parent node, and from the target point Q goal Begin by selecting the parent node Q that connects each node in turn. parent Until the initial point Qstart , as the initial point Q start Move to target point Q goal If the path is correct, the iteration process ends; otherwise, repeat steps two through six to continue expanding the random tree.

[0096] Step Nine: As Figure 7 , 8 As shown, the path is pruned, redundant nodes are removed, and the random tree is trimmed. The target point Q is... goal As the endpoint, the initial point Q start As the starting point, connect the ending point to the starting point; if the line connecting the starting and ending points collides with an obstacle, then the child node Q of the starting point is used as the starting point. child Update the starting point until the line connecting the starting and ending points does not collide with any obstacles. Use the updated starting point as the ending point, and the initial point is Q. start Starting from point Q, repeat the above process until the endpoint coincides with the initial point Q. start The line connecting the obstacles did not collide with them.

[0097] Each node in the path is selected sequentially as the moving node Q. choose First move node Q choose Moving towards its parent node, the translation vector d1 for each move is:

[0098]

[0099] Until the moving node Q choose The connection between the node and its child nodes collides with an obstacle or moves node Q. choose After reaching its parent node, stop moving towards its parent node; then move the node Q. choose Move towards its child nodes, and the translation vector d2 for each move is:

[0100]

[0101] Until the moving node Q choose The connection between the node and its parent node collides with an obstacle or moves node Q. choose Once a node reaches its child node, stop moving towards its child nodes. After moving all nodes in the path, obtain the updated path as the initial point Q. start Move to target point Q goal The final path.

[0102] Step 10: As Figure 9 , Figure 10As shown in the figure, the present invention and the RRT* algorithm were used to perform path planning on complex maps and single-pass narrow road maps, respectively. It can be seen from the figure that the path generated by the present invention has fewer turns and smaller turn curvature than the path generated by the RRT* algorithm. The average time and average path length of the present invention, RRT*, and trimmed RRT* algorithm for 10 path planning operations on complex maps are shown in Table 1.

[0103] Table 1 Comparison of Final Paths

[0104]

[0105]

[0106] As can be seen from Table 1, the final path generated by this invention is superior to the traditional RRT* algorithm in terms of both time and path length.

Claims

1. A robot path planning method based on an improved RRT algorithm, characterized in that: Includes the following steps: Step 1: Obtain environmental information around the robot and establish a coordinate system; environmental information includes the initial point. Q start Obstacles and target points Q goal Location information; initial point Q start Add it to the random tree as the root node; Step 2: Obtain sampling points in the constructed coordinate system Q sample ; Sampling points Q sample azimuth φ and sampling radius r limit The expression is: ； in, a 1 represents the goal-orientation coefficient. a 1> k max - k min ; θ initial As the initial point Q start and target point Q goal The connection and x The included angle formed by the axes; k For random numbers with an expected value of zero, k ∈[ k min , k max ]; k min and k max Each is a preset random number k The minimum and maximum values; a 2 represents the global coefficient for radians; a 3 is the expansion factor; r max The maximum sampling radius; a 4 represents the global step size coefficient; r new For the previous newly added node relative to the initial point Q start The Euclidean distance; a 5 represents the collision coefficient with the sampling. col Related step size random coefficients, a 5= rand - col ; rand A random number within a preset range; col The sampling collision coefficient; Q For nodes on a random tree; esp The maximum growth step size is expressed as follows: ； in, s This is the preset minimum growth step size; x step This is the step size coefficient; If the random tree only has a root node, then the root node is used as the previous newly added node. Based on sampling points Q sample Get random points Q rand ; Step 3: Traverse all nodes in the random tree and calculate the relationship between each node and the random point. Q rand Euclidean distance, obtain the distance to random points Q rand The nearest node is used as the feature node. Q near Based on random points Q rand and feature nodes Q near Get new nodes Q new ; Step 4: If a new node is added Q new With feature nodes Q near If the connecting line collides with an obstacle, repeat steps two and three until a new node is obtained. Q new With feature nodes Q near The connection is free of obstacles, and the newly acquired node will be obtained. Q new Add to the random tree; based on the number of collisions num Update step size coefficient x step and sampling collision coefficient col ; In step four, the step size coefficient x step The update method is as follows: based on the original step size coefficient x step Based on adding a small step size Δ x step tiny step size Δ x step With the number of collisions num The sampling collision coefficient decreases as the number of samples increases. col The update method is as follows: based on the original sampled collision coefficients col Add a small coefficient Δ to the base col The small coefficient Δ col The initial value is 0. After two consecutive collisions, for each additional collision, the small coefficient Δ is increased. col Increase the fixed value until a tiny coefficient Δ is reached. col Reaching the preset maximum value; Step 5: Update the newly added node Q new parent node Q parent ; Step 6: Traverse the existing nodes in the random tree and calculate the distance between each node and the target point. Q goal The Euclidean distance between them; if there exists a node and the target point Q goal The Euclidean distance between them is less than the minimum growth step size. s Then that node is taken as the target point. Q goal Find the parent node and obtain the initial point. Q start Move to target point Q goal If the path is correct, the iteration process ends; otherwise, repeat steps two through five to continue expanding the random tree.

2. The robot path planning method based on the improved RRT algorithm according to claim 1, characterized in that: In step two, if the robot can move in three-dimensional space, sampling points are also constructed. Q sample polar angle θ polar angle θ The expression is: ； in, b 1 represents the goal-orientation coefficient. b 1> k max - k min ; b 2 represents the global coefficient for radians.

3. The robot path planning method based on the improved RRT algorithm according to claim 1, characterized in that: In step five, the initial point is obtained. Q start Move to target point Q goal After determining the path, remove redundant nodes from the path. The method for removing redundant nodes is as follows: With target point Q goal As the endpoint, the initial point Q start As the starting point, connect the ending point to the starting point; if the line connecting the starting and ending points collides with an obstacle, then the child nodes of the starting point are considered as the starting point. Q child Update the starting point until the line connecting the starting point and the end point does not collide with any obstacles; Using the updated starting point as the endpoint, the initial point... Q start Starting from this point, repeat the above process until the endpoint is the same as the initial point. Q start The line connecting the obstacles did not collide with them.

4. The robot path planning method based on the improved RRT algorithm according to claim 3, characterized in that: In step five, after removing redundant nodes, each node in the path is selected sequentially as a moving node. Q choose First move the node Q choose Move towards its parent node, the translation vector for each move. d 1 is: ； Until the moving node Q choose The connection between the node and its child nodes collides with or moves the node due to obstacles. Q choose Once the node reaches its parent node, stop moving towards the parent node; then move the node... Q choose Move to its child nodes, the translation vector for each move. d 2 is: ； Until the moving node Q choose The connection between the node and its parent node collides with an obstacle or moves the node. Q choose Once a child node is reached, stop moving towards its child nodes; after moving all nodes in the path, obtain the updated path as the starting point. Q start Move to target point Q goal The final path.

5. A robot path planning method based on an improved RRT algorithm according to claim 1, characterized in that: In step two, if there is no potential field during the acquisition of sampling points... U req sampling points Q sample As a random point Q rand Conversely, random points Q rand The expression is: ； in, Q req The center point of the potential field; F req Sampling points Q sample The repulsive force it receives.

6. A robot path planning method based on an improved RRT algorithm according to claim 5, characterized in that: The sampling points Q sample Repulsive force received F req The expression is: ； in, k req The potential field gain coefficient; r The radius of the potential field is denoted as . n This is the adjustment factor coefficient.

7. A robot path planning method based on an improved RRT algorithm according to claim 5, characterized in that: The potential field U req The confirmation method is as follows: If the neighboring node set Q neighbor If the number of elements in the brackets exceeds a preset threshold, then a new node will be added. Q new As the center point of the potential field Q req The generation only affects the next random point. Q rand The potential field U req And reassign the maximum sampling radius. r max and distance r new The maximum sampling radius after assignment and distance The expression is: ； ； The maximum sampling radius after assignment and distance The maximum sampling radius of the next sampling node r max and distance r new .

8. A robot path planning method based on an improved RRT algorithm according to claim 1, characterized in that: In step five, the newly added node is updated. Q new parent node Q parent The method is as follows: Add a new node with the center of the circle Q new The radius is the radius of the potential field. r All nodes inside the circle are considered as the set of neighboring nodes. Q neighbor []; Get the set of neighboring nodes Q neighbor From each node in [] to the new node Q new Total cost cost : ； in, Q neighbor For the set of neighboring nodes Q neighbor The neighboring nodes in []. cost neighbor As the initial point Q start Reaching neighboring nodes Q neighbor The cost; With neighboring node set Q neighbor From each node in [] to the new node Q new The node corresponding to the minimum cost is taken as the minimum cost node. Q min If the minimum cost node Q min With new nodes Q new If there are no collisions, then the node with the minimum cost is selected. Q min As a new node Q new Parent node.

9. A robot path planning method based on an improved RRT algorithm according to claim 1, characterized in that: In step three, the newly added node is obtained. Q new The method is as follows: ； in, For random points Q rand With feature nodes Q near The Euclidean distance between them.

Images