Mechanical arm path planning method and device, medium and equipment

Abstract

The invention discloses a mechanical arm path planning method and device, a medium and equipment, and relates to the technical field of automatic control. The method comprises the steps that position parameters of a to-be-grabbed target are acquired; improving a bidirectional fast expansion random tree algorithm by using an artificial potential field to obtain an improved bidirectional fast expansion random tree algorithm; the position of an end effector of a mechanical arm is used as a starting node, the position parameter of a target to be grabbed is used as a target node, and a moving path is generated based on an improved two-way fast expansion random tree algorithm. According to the scheme, a high-quality path can be generated efficiently.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic control, and in particular to a robot arm path planning method, device, medium and equipment. Background Art

[0002] The robot arm path planning aims to plan a path from the starting point to the target node for the end effector in a fixed workspace. Driven by the robot arm, it avoids obstacles in its workspace and then moves from the starting point to the terminal node. The key to this process is to find a suitable and high-performance algorithm to implement it.

[0003] The algorithms used in existing robot path planning methods are mainly the rapidly-exploring Random Tree (RRT) algorithm, the rapidly-exploring RandomTree* (RRT*) algorithm and the bidirectional rapidly-exploring RandomTrees* (BI-RRT*) algorithm.

[0004] The RRT algorithm is a path planning algorithm that constructs a search tree through random sampling. It is widely used in fields such as robot path planning. The basic idea of ​​RRT is to start from the starting node, randomly sample points in the state space, and expand to the tree, thereby gradually covering the entire space. Each expansion moves towards the random node closest to a node in the current tree, forming a progressively growing tree until the tree is connected to the target node or target area. The specific steps of the RRT algorithm are: Step 1: Given a map space M, a starting node xstart and a target node xgoal, initialize the tree V={xstart}; Step 2: Get xrand by random sampling in space; Step 3: Starting from the node xnear closest to xrand in the tree V, expand towards xrand with a step size S to get a new node xnew; Step 4: Perform collision detection. If there is no collision, add xnear to the tree V. If there is a collision, re-execute steps 2 and 3; Step 5: Repeat steps 2 to 4 until the distance between a certain xnew and xgoal is less than the preset threshold, stop expanding and generate a path. The advantages of the RRT algorithm lie in its randomness and efficiency, especially in the ability to quickly explore feasible paths in high-dimensional complex environments.

[0005] The RRT* algorithm is an optimized version of the traditional RRT algorithm, which aims to improve the path quality by introducing a reconnection strategy and approach the global optimal solution more efficiently. Similar to the RRT algorithm, the RRT* algorithm still expands the tree through random sampling, but during the expansion process, it not only searches for the nearest node, but also tries to connect the better paths in the existing nodes through a rewiring strategy, thereby gradually optimizing the generated path. The basic steps of RRT* are as follows: Step 1: Given a map space M, a starting node xstart, a target node xgoal, and an initialized tree V={xstart}; Step 2: Get xrand by random sampling in the space; Step 3: Start from the node xnear closest to xrand in the tree V, and expand towards xrand with a step length S to get a new node xnew; Step 4: Perform collision detection. If there is no collision, add xnear to the tree V. If there is a collision, re-execute steps 2 and 3; Step 5: If xnew is generated, calculate the distance d from xnew to xstart through each surrounding node, select the node with the shortest d as the parent node of xnew, and connect xnew and its parent node; Step 6: Repeat steps 2 to 5 until the distance between a certain xnew and xgoal is less than the preset threshold, stop expanding, and generate a path. The RRT* algorithm optimizes the path quality by introducing a reconnection strategy and gradually approaches the global optimal solution, which improves the path quality problem of traditional RRT.

[0006] The BI-RRT* algorithm is an improvement on the RRT* algorithm. The algorithm constructs two RRT* trees from xstart and xgoal at the same time, one expanding from xstart to xgoal, and the other exploring from xgoal to xstart. The expansion process of the two trees is the same as that of the RRT* algorithm. During the expansion process, the two trees continue to approach each other until the distance between a pair of nodes between them is less than the preset threshold. At this point, the algorithm stops expanding and connects the two trees to generate a valid path from xstart to xgoal. BI-RRT* improves the efficiency of path search through a bidirectional expansion strategy, can quickly explore complex spaces, and is particularly suitable for high-dimensional and complex environments. It is one of the best performing path planning algorithms in the current RRT algorithm series.

[0007] However, although the RRT algorithm can find a feasible path, the generated path is usually of poor quality and may not be smooth enough. More importantly, the RRT algorithm may take a very long time to find the target in some cases, especially in complex environments, where the expansion of the tree may require a lot of invalid sampling and expansion. Due to its random sampling and aimless exploration methods, the RRT* algorithm still has problems such as poor path quality, long calculation time and low efficiency in complex environments. The BI-RRT* algorithm lacks the use of task information in the path planning process, and still randomly expands aimlessly, so the quality of the path it generates is still unsatisfactory, which may require more time and computing resources. Therefore, the existing robot arm path planning methods cannot efficiently generate high-quality paths. Summary of the invention

[0008] Based on this, it is necessary to provide a robot arm path planning method, device, medium and equipment to address the above technical problems.

[0009] The present invention adopts the following technical solutions: In a first aspect, the present invention provides a robot arm path planning method, the method comprising: Get the position parameters of the target to be captured; The artificial potential field is used to improve the bidirectional rapid expansion random tree algorithm, and an improved bidirectional rapid expansion random tree algorithm is obtained; Taking the position of the end effector of the robot arm as the starting node and the position parameter of the target to be grasped as the target node, generating a moving path based on the improved bidirectional rapid expansion random tree algorithm; The improvement of the bidirectional fast expanding random tree algorithm by using an artificial potential field specifically includes: For the i-th node generated by the bidirectional fast expanding random tree, a node having the shortest Euclidean distance to the i-th node is determined as a neighboring node; Calculate the attraction of the target node and the random node on the neighboring node, and calculate the repulsion between the neighboring node and the obstacle associated with the i-th node, wherein the random node is a node randomly sampled in the process of expanding the node by the bidirectional fast expansion random tree algorithm; The expansion direction of the neighboring nodes is determined according to the attractive force and the repulsive force, and node expansion is performed according to the expansion direction of the neighboring nodes to obtain the (i+1)th node.

[0010] Furthermore, the calculation formula of the attraction is: + ; in, For attraction,k a is the coefficient of the attraction field, is the location of the neighboring node, is the location of the target node, is the position of the random node, d ( x near , x goal )express x near and x goal The Euclidean distance between express x near and x rand The Euclidean distance between The calculation formula of the repulsive force is: ; in, is the repulsive force, is the coefficient of the repulsive force field, d 0 is the maximum range of the repulsive force of the obstacle, is the random node with the shortest Euclidean distance to the i-th node, x obs is the center position of the obstacle, d ( x near , x obs )express x near and x obs The Euclidean distance between .

[0011] Furthermore, the expressions of the attractive force field and the repulsive force field are respectively: ; ; ; in, is the attraction field of the target node to the neighboring nodes, is the attraction field of the random node to the neighboring node, k a is the coefficient of the attraction field, is the location of the neighboring node, is the position of the obstacle center, is the location of the target node, is the position of the random node, d ( x near , x goal )express x near and x goal The Euclidean distance between is the repulsive force field, is the coefficient of the repulsive force field, d 0 is the maximum range of the obstacle repulsion, Represents the Euclidean distance between the neighboring node and the obstacle center.

[0012] Furthermore, after obtaining the improved bidirectional fast expanding random tree algorithm, it also includes: Introducing a bias strategy into the sampling operation of the improved bidirectional fast expanding random tree algorithm; The expression of the bias strategy is: ; in, ' represents the current sampling node, is the target node, is a random node, is the preset bias probability, rand Represents the random number generated in the sampling operation.

[0013] Further, after generating the moving path based on the improved bidirectional rapid expansion random tree algorithm, the method further includes: Optimizing the moving path by using a greedy strategy to obtain an optimized moving path; The steps of the greedy strategy include: Connecting the target node with the starting node to obtain a straight line path; Performing a collision detection on the straight path, and if there is no obstacle on the straight path, determining the straight path as the optimized moving path; If there is an obstacle on the straight path, each parent node is detected step by step starting from the target node to determine J reserved paths, and the J reserved paths are connected in sequence to obtain the optimized moving path; The jth reserved path is determined as follows: starting from the target node, each parent node is detected step by step until the parent node with no obstacles in the straight path between the j-1th optimization node and the target node is detected as the jth optimization node, and the straight path between the j-1th optimization node and the j-1th optimization node is the jth reserved path, J-1≥j≥1, when j=1, the j-1th optimization node is the starting node, J≥2; The Jth retained path is a straight line path between the target node and the J-1th optimized node, and there are no obstacles on the straight line path between the target node and the J-1th optimized node.

[0014] In a second aspect, the present invention provides a robot arm path planning device, comprising: An acquisition module is used to obtain the position parameters of the target to be captured; An improvement module is used to improve the bidirectional rapid expansion random tree algorithm by using an artificial potential field to obtain an improved bidirectional rapid expansion random tree algorithm; A path generation module, used to generate a moving path based on the improved bidirectional rapid expansion random tree algorithm, taking the position of the end effector of the robot arm as the starting node and the position parameter of the target to be grasped as the target node; The improvement of the bidirectional fast expanding random tree algorithm by using an artificial potential field specifically includes: For the i-th node generated by the bidirectional fast expanding random tree, a node having the shortest Euclidean distance to the i-th node is determined as a neighboring node; Calculate the attraction of the target node and the random node on the neighboring node, and calculate the repulsion between the neighboring node and the obstacle associated with the i-th node, wherein the random node is a node randomly sampled in the process of expanding the node by the bidirectional fast expansion random tree algorithm; The expansion direction of the neighboring nodes is determined according to the attractive force and the repulsive force, and node expansion is performed according to the expansion direction of the neighboring nodes to obtain the (i+1)th node.

[0015] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the robot arm path planning method is implemented.

[0016] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the robot arm path planning method when executing the program.

[0017] The at least one technical solution adopted by the present invention can achieve the following beneficial effects: the present invention improves the bidirectional rapid expansion random tree algorithm by using an artificial potential field, calculates the attractive force of the target node and the random node on the adjacent node, and calculates the repulsive force between the adjacent node and the obstacle; determines the expansion direction of the adjacent node according to the attractive force and the repulsive force, and expands the node according to the expansion direction of the adjacent node, and can combine the efficient search characteristics of the bidirectional rapid expansion random tree and the guiding role of the artificial potential field in the path generation process, optimize the quality of path generation and the search efficiency, ensure the efficiency of searching for nodes in the path planning process, thereby approaching the target more directly, while avoiding the interference of obstacles, and achieving efficient generation of high-quality paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0019] Figure 1 A schematic diagram of a process flow of a robot arm path planning method provided by the present invention; Figure 2 A force analysis diagram of adjacent nodes in an artificial potential field provided by the present invention; Figure 3 A schematic diagram of optimizing a moving path using a greedy strategy provided by the present invention; Figure 4 A schematic diagram of a robot arm path planning device provided by the present invention; Figure 5 A schematic diagram of a computer device for implementing a robot arm path planning method provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] The server mentioned in the present invention can be a server set on a business platform, or a device such as a desktop computer, a laptop computer, etc. that can execute the solution of the present invention. For the convenience of description, the following description is only based on the server as the execution subject. The following is a detailed description of the technical solutions provided by various embodiments of the present invention in conjunction with the accompanying drawings.

[0022] refer to Figure 1, a robot arm path planning method in the present invention specifically comprises the following steps: S10: Obtain the position parameters of the target to be grasped.

[0023] In this embodiment, the target to be grasped refers to an object that needs to be grasped by the robot arm, and the position parameter refers to the three-dimensional coordinates of the position of the target to be grasped.

[0024] S20: The bidirectional rapidly expanding random tree algorithm is improved by using an artificial potential field to obtain an improved bidirectional rapidly expanding random tree algorithm.

[0025] In this embodiment, an artificial potential field (APF) is used to improve the bidirectional rapid expanding random tree algorithm to obtain an improved bidirectional rapid expanding random tree algorithm.

[0026] Among them, the artificial potential field is used to improve the bidirectional fast expansion random tree algorithm, including: For the i-th node generated by the bidirectional rapid expansion random tree, the node with the shortest Euclidean distance to the i-th node is determined as a neighboring node.

[0027] The attraction of the target node and the random node on the neighboring nodes is calculated, and the repulsion between the neighboring nodes and the obstacles associated with the i-th node is calculated. The random nodes are nodes randomly sampled in the process of expanding nodes by the bidirectional fast expansion random tree algorithm.

[0028] The expansion direction of the neighboring nodes is determined according to the attraction and repulsion forces, and the nodes are expanded according to the expansion direction of the neighboring nodes to obtain the i+1th node.

[0029] In this embodiment, reference Figure 2 , is the neighboring node, Obstacle For obstacles ， is a random node, 1 is the attraction of a random node to its neighboring nodes, 2 is the attraction of the target node to the neighboring nodes, is the repulsive force of the obstacle on the adjacent nodes, F total for 1. 2 The vector sum of .

[0030] In this embodiment, the calculation formula of attraction is: + (1) in, For attraction, k a is the coefficient of the attraction field, is the location of the neighboring node, is the location of the target node, is the position of a random node, d ( x near , x goal )express x near and x goal The Euclidean distance between express x near and x rand The Euclidean distance between . is the attraction of the target node to the neighboring nodes, is the attraction of a random node to its neighboring nodes.

[0031] The formula for calculating the repulsive force is: (2) in, is the repulsive force, is the coefficient of the repulsive force field, d 0 is the maximum range of the repulsive force of the obstacle, is the random node with the shortest Euclidean distance to the i-th node, x obs is the center position of the obstacle, d ( x near , x obs )express x near and x obs The Euclidean distance between .

[0032] Specifically, the expressions of the attractive force field and the repulsive force field are: (3) (4) (5) in, is the attraction field, k a is the coefficient of the attraction field, is the location of the neighboring node, is the position of the obstacle center, is the location of the target node, is the position of a random node, d ( x near , x goal )express x near and x goal The Euclidean distance between is the repulsive force field, is the coefficient of the repulsive force field, d 0 is the maximum range of the obstacle repulsion, Indicates the Euclidean distance between the neighboring node and the obstacle center.

[0033] S30: Taking the position of the end effector of the robot arm as the starting node and the position parameter of the target to be grasped as the target node, a moving path is generated based on an improved bidirectional fast expanding random tree algorithm.

[0034] In this embodiment, the position of the end effector of the robot arm is used as the starting node, and the position parameters of the target to be grasped are used as the target node. A moving path is generated based on an improved bidirectional rapid expansion random tree algorithm, so that the end effector of the robot arm moves along the generated moving path to the target to be grasped, and grasps the target to be grasped.

[0035] based on Figure 1 A robot arm path planning method shown in the invention improves the bidirectional rapid expansion random tree algorithm by using an artificial potential field, calculates the attractive force of the target node and the random node on the adjacent node, and calculates the repulsive force between the adjacent node and the obstacle; determines the expansion direction of the adjacent node according to the attractive force and the repulsive force, and expands the node according to the expansion direction of the adjacent node, and can combine the efficient search characteristics of the bidirectional rapid expansion random tree and the guiding role of the artificial potential field in the path generation process, optimize the quality of path generation and the search efficiency, ensure the efficiency of searching for nodes in the path planning process, so as to approach the target more directly, avoid the interference of obstacles, and realize the efficient generation of high-quality paths.

[0036] When applying a robot arm path planning method provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0037] In addition, in one or more embodiments of the present invention, the bidirectional fast expanding random tree algorithm is improved by using an artificial potential field, further comprising: A bias strategy is introduced into the sampling operation of the bidirectional rapidly expanding random tree algorithm.

[0038] The expression of the bias strategy is: (5) in, ' represents the current sampling node, is the target node, is a random node, is the preset bias probability, rand Represents the random number generated in the sampling operation.

[0039] In this embodiment, the specific steps of introducing a bias strategy into the sampling operation of the bidirectional rapid expansion random tree algorithm include: Step 1: Define two starting nodes x in the workspace star1 and x star2 , and two target nodes x goal1 and x goal2 Initialize two random trees T1 and T2, T1 and T2 contain the starting node x star1 and x star2 .

[0040] Step 2: randomly generate a sampling node x in the entire workspace rand According to the set bias probability P bias , x rand Biased towards one of the target nodes x goal1 or x goal2 , to increase the chances of expansion in the target direction.

[0041] Step 3: For each random tree, find the distance from the current sampling node x rand The nearest node x near .

[0042] Step 4: For each nearest node x near Calculate the force field, which includes: 1) Attraction: target node x goal x near Apply attraction F ai , sampling node x rand Also for x near Apply attraction F aj .

[0043] 2) Repulsive force: at x near Obstacles within a certain range around x near Apply repulsive force F ri .

[0044] 3) Resultant force: Calculate the vector sum of the attractive and repulsive forces to get x nearThe direction of the resultant force is determined as x near The final direction of expansion F t .

[0045] Step 5: Expand the node according to the resultant force F t Direction, from x near Expand in this direction with a fixed step size s to generate a new node x new .

[0046] Step 6: Generate the new node x new Perform collision detection.

[0047] If the test passes, x new x near Add the parent node to the corresponding random tree T1 or T2. new Search the nodes in the existing tree within a certain radius r, and select the nodes from the starting node to x. new The node x with the lowest path cost min , and x new The parent node is updated to x min , in order to optimize the movement path.

[0048] If the test fails, the current expansion is abandoned and the process returns to step 2 to resample and expand the nodes.

[0049] Step 7, check connections and iterations: 1) Repeat steps 2 to 6, checking in each iteration whether the two random trees T1 and T2 have been successfully connected to form a complete path.

[0050] 2) Determine whether the preset maximum number of iterations has been reached.

[0051] Step 8, Generate and optimize the path: 1) Successful connection: If two trees are successfully connected, stop sampling and generate a preliminary path from the starting point to the target point. Then, the path is optimized through a greedy strategy to improve the overall quality of the path.

[0052] 2) Reaching the maximum number of iterations: If two trees cannot be connected within the maximum number of iterations, the algorithm terminates.

[0053] This embodiment introduces a bias strategy for the target node during the sampling process, so that the probability of the target node being sampled is much higher than that of other random nodes, thereby increasing the coverage frequency of the target area, shortening the tree expansion path, and reducing the time of invalid expansion, thereby reducing unnecessary exploration in the search space and improving the convergence speed and performance of the improved bidirectional fast expansion random tree algorithm in a practical environment.

[0054] In addition, in one or more embodiments of the present invention, after generating the moving path based on the improved bidirectional fast expanding random tree algorithm, the method further includes: The moving path is optimized through the greedy strategy to obtain the optimized moving path.

[0055] Among them, the steps of the greedy strategy include: Connect the target node with the starting node to obtain a straight line path.

[0056] A collision detection is performed on the straight line path. If there is no obstacle on the straight line path, the straight line path is determined to be the optimized moving path.

[0057] If there are obstacles on the straight path, each parent node is detected step by step starting from the target node, J reserved paths are determined, and the J reserved paths are connected in sequence to obtain the optimized moving path.

[0058] Among them, the jth retained path is determined as follows: starting from the target node, each parent node is detected step by step until the parent node whose straight path with the j-1th optimization node has no obstacles is detected as the jth optimization node, and the straight path between the jth optimization node and the j-1th optimization node is the jth retained path, J-1≥j≥1, when j=1, the j-1th optimization node is the starting node, J≥2.

[0059] The Jth retained path is a straight line path between the target node and the J-1th optimized node, and there are no obstacles on the straight line path between the target node and the J-1th optimized node.

[0060] In this embodiment, the parent node refers to the upper level node, and the child node refers to the lower level node. The relationship between two adjacent nodes is: the node close to the starting node is the parent node, and the node close to the target node is the child node. The optimized node is the node between the starting node and the target node, and there are no obstacles in the straight path between any two optimized nodes. For example, if three reserved paths are determined between the starting node and the target node, there are two optimized nodes between the starting node and the target node, wherein the first reserved path and the second reserved path are connected through the first optimized node, and the second reserved path and the third reserved path are connected through the second optimized node.

[0061] In this embodiment, the reserved path refers to a straight path without any obstacles. Figure 3 , respectively showing the original moving path, the first optimization process of the moving path, the second optimization process of the moving path, and the optimized moving path. start is the starting node, X goal is the target node, and Obstacle is the obstacle. Figure 3It can be seen that by continuously backtracking and updating the moving path, the greedy strategy effectively simplifies the moving path, while ensuring that the moving path is efficient and direct from the starting point to the target point, significantly reducing the collision-free nodes.

[0062] This embodiment introduces a greedy strategy to the improved bidirectional rapid expansion random tree algorithm, continuously backtracks and updates the path during the generation of the mobile path, effectively simplifies the path, and at the same time ensures that the path is efficient and direct from the starting point to the target point, significantly reduces the collision-free nodes, effectively reduces the number of path turns and the path cost, and makes the quality of the generated mobile path higher.

[0063] The above is a robot arm path planning method provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding robot arm path planning device, such as Figure 4 As shown, including: The acquisition module is used to obtain the position parameters of the target to be captured.

[0064] The improvement module is used to improve the bidirectional rapid expansion random tree algorithm by using an artificial potential field to obtain an improved bidirectional rapid expansion random tree algorithm.

[0065] The path generation module is used to generate a moving path based on an improved bidirectional rapid expansion random tree algorithm, taking the position of the end effector of the robot arm as the starting node and the position parameters of the target to be grasped as the target node.

[0066] Among them, the artificial potential field is used to improve the bidirectional fast expansion random tree algorithm, including: For the i-th node generated by the bidirectional rapid expansion random tree, the node with the shortest Euclidean distance to the i-th node is determined as a neighboring node.

[0067] The attraction of the target node and the random node on the neighboring nodes is calculated, and the repulsion between the neighboring nodes and the obstacles associated with the i-th node is calculated. The random nodes are nodes randomly sampled in the process of expanding nodes by the bidirectional fast expansion random tree algorithm.

[0068] The expansion direction of the neighboring nodes is determined according to the attraction and repulsion forces, and the nodes are expanded according to the expansion direction of the neighboring nodes to obtain the i+1th node.

[0069] For the specific definition of a robot path planning device, please refer to the definition of a robot path planning method above, which will not be repeated here. The various modules in the robot path planning device can be implemented in whole or in part by software, hardware and a combination thereof. The modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0070] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the Figure 1 A robot arm path planning method is provided.

[0071] The present invention also provides Figure 5 The structural diagram of the computer device shown in FIG. Figure 5 As shown, at the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the Figure 1 A robot arm path planning method is provided.

[0072] A person of ordinary skill in the art can understand that all or part of the processes in the embodiment method can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the methods described. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0073] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A robot arm path planning method, characterized in that: include: Get the position parameters of the target to be captured; The artificial potential field is used to improve the bidirectional rapid expansion random tree algorithm, and an improved bidirectional rapid expansion random tree algorithm is obtained; Taking the position of the end effector of the robot arm as the starting node and the position parameter of the target to be grasped as the target node, generating a moving path based on the improved bidirectional rapid expansion random tree algorithm; The improvement of the bidirectional fast expanding random tree algorithm by using an artificial potential field specifically includes: For the i-th node generated by the bidirectional fast expanding random tree, a node having the shortest Euclidean distance to the i-th node is determined as a neighboring node; Calculate the attraction of the target node and the random node on the neighboring node, and calculate the repulsion between the neighboring node and the obstacle associated with the i-th node, wherein the random node is a node randomly sampled in the process of expanding the node by the bidirectional fast expansion random tree algorithm; The expansion direction of the neighboring nodes is determined according to the attractive force and the repulsive force, and node expansion is performed according to the expansion direction of the neighboring nodes to obtain the (i+1)th node.

2. A robot arm path planning method as claimed in claim 1, characterized in that: The calculation formula of the attraction is: + ； in, For the total attraction, k a is the coefficient of the attraction field, is the location of the neighboring node, is the location of the target node, is the position of the random node, d ( x near , x goal )express x near and x goal The Euclidean distance between express x near and x rand The Euclidean distance between The calculation formula of the repulsive force is: ； in, is the repulsive force, k r is the coefficient of the repulsive force field, d 0 is the maximum range of the repulsive force of the obstacle, x obs is the center position of the obstacle, d ( x near , x obs )express x near and x obs The Euclidean distance between .

3. A robot arm path planning method as claimed in claim 2, characterized in that: The expressions of the attractive force field and the repulsive force field are respectively: ； ； ； in, is the attraction field of the target node to the neighboring nodes, is the attraction field of the random node to the neighboring node, k a is the coefficient of the attraction field, is the location of the neighboring node, is the position of the obstacle center, is the location of the target node, is the position of the random node, d ( x near , x goal )express x near and x goal The Euclidean distance between is the repulsive force field, is the coefficient of the repulsive force field, d 0 is the maximum range of the obstacle repulsion force, Represents the Euclidean distance between the neighboring node and the obstacle center.

4. A robot arm path planning method as claimed in claim 1, characterized in that: The improvement of the bidirectional fast expansion random tree algorithm by using an artificial potential field also includes: Introducing a bias strategy into the sampling operation of the bidirectional rapidly expanding random tree algorithm; The expression of the bias strategy is: ； in, ' represents the current sampling node, is the target node, is a random node, is the preset bias probability, rand Represents the random number generated in the sampling operation.

5. A robot arm path planning method as claimed in claim 1, characterized in that: After generating the moving path based on the improved bidirectional rapid expansion random tree algorithm, the method further includes: Optimizing the moving path by using a greedy strategy to obtain an optimized moving path; The steps of the greedy strategy include: Connecting the target node with the starting node to obtain a straight line path; Performing a collision detection on the straight path, and if there is no obstacle on the straight path, determining the straight path as the optimized moving path; If there is an obstacle on the straight path, each parent node is detected step by step starting from the target node to determine J reserved paths, and the J reserved paths are connected in sequence to obtain the optimized moving path; The jth reserved path is determined as follows: starting from the target node, each parent node is detected step by step until the parent node with no obstacles in the straight path between the j-1th optimized node and the j-1th optimized node is detected as the j-th optimized node, and the straight path between the j-1th optimized node and the j-1th optimized node is the j-th reserved path, J-1≥j≥1, when j=1, the j-1th optimized node is the starting node, J≥2; The Jth retained path is a straight line path between the target node and the J-1th optimized node, and there are no obstacles on the straight line path between the target node and the J-1th optimized node.

6. A robot arm path planning device, characterized in that: include: An acquisition module is used to obtain the position parameters of the target to be captured; An improvement module is used to improve the bidirectional rapid expansion random tree algorithm by using an artificial potential field to obtain an improved bidirectional rapid expansion random tree algorithm; A path generation module, used to generate a moving path based on the improved bidirectional rapid expansion random tree algorithm, taking the position of the end effector of the robot arm as the starting node and the position parameter of the target to be grasped as the target node; The improvement of the bidirectional fast expanding random tree algorithm by using an artificial potential field specifically includes: For the i-th node generated by the bidirectional fast expanding random tree, a node having the shortest Euclidean distance to the i-th node is determined as a neighboring node; Calculate the attraction of the target node and the random node on the neighboring node, and calculate the repulsion between the neighboring node and the obstacle associated with the i-th node, wherein the random node is a node randomly sampled in the process of expanding the node by the bidirectional fast expansion random tree algorithm; The expansion direction of the neighboring nodes is determined according to the attractive force and the repulsive force, and node expansion is performed according to the expansion direction of the neighboring nodes to obtain the (i+1)th node.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method for robot arm path planning according to any one of claims 1 to 5 is implemented.

8. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a robot arm path planning method as claimed in any one of claims 1 to 5 is implemented.

Images