Ultrasonic enhanced robot obstacle avoidance path planning method and device

By building industrial robot and CNC machining environment models, using adaptive expansion step size and Gaussian sampling strategy to optimize the path, the problems of manual teaching dependence and poor path quality in the existing technology are solved, and efficient collision-free in complex environments are achieved Path planning.

CN118999560APending Publication Date: 2024-11-22EAST CHINA UNIV OF SCI & TECH
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411052021.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In the prior art, ultrasonic enhanced robot obstacle avoidance path planning relies on manual teaching, which makes it time-consuming and poor path quality in a composite CNC machining environment, making it difficult to effectively improve path planning efficiency and reduce collision occurrence.

Method used

By building industrial robot models and complex CNC machining environment models, the starting point and end point are used as nodes on the tree, and the new nodes are gradually expanded using the adaptive expansion step size, and the step size is updated according to the congestion of the spatial environment during the expansion process. Combined with Gaussian sampling strategy, multiple iterative optimizations are performed on the collision-free path to obtain the optimal path.

Benefits of technology

It improves path planning efficiency, reduces the possibility of collision occurrence, can quickly find the optimal collision-free path in complex environments, and improves the automation and production efficiency of industrial ultrasonic enhanced robot obstacle avoidance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118999560A_ABST
    Figure CN118999560A_ABST
Patent Text Reader

Abstract

The invention discloses an ultrasonic enhanced robot obstacle avoidance path planning method and device, and belongs to the field of robot simulation. The method comprises the following steps: constructing an industrial robot model and a complex numerical control machining environment model; taking the starting point and the ending point as nodes on a tree, and gradually expanding new nodes on the tree from the starting point according to a self-adaptive expansion step length by using the industrial robot model and the processing environment model so as to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the crowdedness degree of the space environment; and utilizing a Gaussian sampling strategy to carry out multiple times of iterative optimization on the collision-free path so as to obtain an optimal collision-free path. According to the invention, the path planning efficiency can be improved and collision can be reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of robot simulation technology, and in particular to an ultrasonic enhanced robot obstacle avoidance path planning method and device. Background Art

[0002] Industrial ultrasonic enhanced robot obstacle avoidance path planning is essentially to find a collision-free path from the starting point to the end point. The shorter the path length, the shorter the running time and the higher the operating efficiency. Ultrasonic enhanced robot obstacle avoidance path planning in related technologies is mainly completed through manual teaching. However, the manual teaching method is highly dependent on the experience of engineers. In the face of complex CNC machining environments, manual planning is not only time-consuming, but also has poor path quality. Summary of the invention

[0003] The present invention provides a method and device for planning an obstacle avoidance path for an ultrasonic enhanced robot. The technical solution is as follows:

[0004] On the one hand, a method for obstacle avoidance path planning of an ultrasound enhanced robot is provided, the method comprising:

[0005] Build industrial robot models and complex CNC machining environment models;

[0006] The starting point and the end point are used as nodes on the tree, and the industrial robot model and the processing environment model are used to gradually expand new nodes on the tree from the starting point according to the adaptive expansion step length to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the congestion degree of the spatial environment;

[0007] The collision-free path is optimized multiple times by using Gaussian sampling strategy to obtain the optimal collision-free path.

[0008] On the other hand, an ultrasonic enhanced robot obstacle avoidance path planning device is provided, the device comprising:

[0009] Construction unit, used to build industrial robot models and complex CNC machining environment models;

[0010] A node expansion unit, used to use the starting point and the ending point as nodes on the tree, and to use the industrial robot model and the processing environment model to gradually expand new nodes on the tree from the starting point according to the adaptive expansion step length to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the congestion degree of the spatial environment;

[0011] The path optimization unit is used to perform multiple iterative optimizations on the collision-free path using a Gaussian sampling strategy to obtain an optimal collision-free path.

[0012] On the other hand, a computer device is provided, which includes a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to implement the steps of the above-mentioned ultrasonic enhanced robot obstacle avoidance path planning method.

[0013] On the other hand, a computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned ultrasonic enhanced robot obstacle avoidance path planning method are implemented.

[0014] On the other hand, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the above-mentioned ultrasound enhanced robot obstacle avoidance path planning method.

[0015] The technical solution provided by the present invention can at least bring the following beneficial effects:

[0016] In the process of path planning, new nodes need to be continuously expanded to obtain a collision-free path connecting the starting point and the end point. When the CNC machining environment is very complex, the degree of congestion in the space environment where different expansion nodes are located is different. By adaptively updating the expansion step size according to the degree of congestion in the space environment, new nodes are expanded using the adaptive expansion step size. After obtaining the collision-free path, the collision-free path is iteratively optimized multiple times using the Gaussian sampling strategy, so that the optimal collision-free path can be obtained. This solution can improve the efficiency of path planning and reduce the occurrence of collisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is a flow chart of an obstacle avoidance path planning method for an ultrasonic enhanced robot provided by an embodiment of the present invention;

[0019] Figure 2 is a detection step diagram of a GJK algorithm provided by an embodiment of the present invention;

[0020] Figure 3 A schematic diagram of a collision-free path obtained by using an adaptive expansion step size is provided in one embodiment of the present invention;

[0021] Figure 4 It is a schematic diagram of a heuristic potential field provided by an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of a sampling range formed by a super ellipsoid provided by an embodiment of the present invention;

[0023] Figure 6 It is a schematic diagram of obtaining collision-free path search results in different complex environments for a two-dimensional scene provided by an embodiment of the present invention;

[0024] Figure 7 It is a schematic diagram of obtaining collision-free path search results in different complex environments of a three-dimensional scene provided by an embodiment of the present invention;

[0025] Figure 8 It is a scene modeling diagram of ultrasonic enhanced robot composite CNC machining provided by an embodiment of the present invention;

[0026] Fig. 9 It is a structural diagram of an ultrasonic enhanced robot obstacle avoidance path planning device provided by one embodiment of the present invention;

[0027] Fig.10 It is a hardware architecture diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in 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.

[0029] Please refer to Figure 1 , an embodiment of the present invention provides an ultrasonic enhanced robot obstacle avoidance path planning method, the method comprising:

[0030] Step 100, constructing an industrial robot model and a complex CNC machining environment model;

[0031] Step 102, taking the starting point and the ending point as nodes on the tree, using the industrial robot model and the processing environment model, gradually expanding new nodes on the tree from the starting point according to the adaptive expansion step length to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the congestion degree of the spatial environment;

[0032] Step 104 , performing multiple iterations of optimization on the collision-free path using a Gaussian sampling strategy to obtain an optimal collision-free path.

[0033] In the embodiment of the present invention, in the process of path planning, new nodes need to be continuously expanded to obtain a collision-free path connecting the starting point and the end point. When the numerical control machining environment is very complex, the degree of congestion of the spatial environment where different expansion nodes are located is different. By adaptively updating the expansion step size according to the degree of congestion of the spatial environment, the new node is expanded using the adaptive expansion step size. After obtaining the collision-free path, the collision-free path is iteratively optimized multiple times using the Gaussian sampling strategy, so that the optimal collision-free path can be obtained. This solution can improve the efficiency of path planning and reduce the occurrence of collisions.

[0034] Described below Figure 1 How the various steps are performed.

[0035] First, for step 100, an industrial robot model and a complex CNC machining environment model are constructed.

[0036] Before performing obstacle avoidance path planning for industrial ultrasonic enhanced robots, collision detection must be performed between the industrial robot and the working environment to help the industrial robot successfully complete obstacle avoidance path planning in a complex working environment and avoid collision accidents. Therefore, it is necessary to analyze the working environment of the industrial robot, reconstruct and export collision detection information points based on the environment layout. However, in the actual processing environment, obstacles often have different shapes, and obstacles and industrial robots need to be abstracted to improve the algorithm operation rate.

[0037] In an embodiment of the present invention, the industrial robot model can be constructed by using a bounding box. Since a directed bounding box is close to the real shape of the industrial robot entity, it is preferred to use a directed bounding box to construct the industrial robot model. It should be noted that the industrial robot model constructed by using a directed bounding box can be obtained from relevant literature and will not be described in detail here.

[0038] Considering the large amount of data and large amount of calculation in the CNC machining environment of actual application, in order to quickly query nodes and reduce memory requirements, in the embodiment of the present invention, an octree algorithm based on hash index is used to construct a complex CNC machining environment model. The constructed machining environment model is a voxelized model. The octree is used to store voxels, which can effectively reduce the high memory requirements of voxels and make it easier to discretize the model. It can effectively improve the voxelization inside the object, and the hash index method can be used to more quickly determine the collision area. Among them, the hash mapping formula is as shown:

[0039] A=n 2 *(z-1)+n*(x-1)+y

[0040] hashKey=floor(m*K*A-floor(K*A))

[0041] Among them, A is the relative position of the voxel block in the world coordinate system, K is the golden ratio, m is the magnification coefficient, n is the number of points in the model surface point cloud, (x, y, z) is the coordinate of a point in the model surface point cloud, and hashKey is the hash value after hash mapping of the point (x, y, z).

[0042] Then, for step 102, the starting point and the end point are used as nodes on the tree, and the industrial robot model and the processing environment model are used to gradually expand new nodes on the tree starting from the starting point according to the adaptive expansion step size to obtain a collision-free path.

[0043] A collision-free path is a multi-node path connecting the starting point to the end point. The path needs to ensure that no collision occurs during the path and the path length can be as short as possible.

[0044] Whenever a sampling node is obtained by sampling the current node, a collision detection needs to be performed on the sampling node to ensure that the newly expanded node is a sampling node that will not collide. In an embodiment of the present invention, the collision detection method may include rough collision detection and precise collision detection. First, an intersection detection is performed based on the AABB bounding box. If there is an intersection, the collision is further calculated between the directed bounding box tree of the industrial robot composed of the actual scene judgment and the octree composed of the voxel method. If there is a collision, finally, a precise collision detection between points and faces is performed based on the basic primitive method.

[0045] The Separating Axis Theorem (SAT) is a common directed bounding box intersection detection algorithm. Its core idea is that if two convex polygons do not collide, a separation line can be found on a certain axis so that the projections of the two polygons on the axis do not overlap. However, when dealing with collision detection between a large number of convex polygons, the SAT algorithm may face performance issues because it needs to detect between each pair of convex polygons, and the complexity will increase exponentially with the increase in the number of cuboids.

[0046] Based on this, the embodiment of the present invention adopts the GJK algorithm for collision detection. As an iterative algorithm with nearly linear complexity, the GJK algorithm converges very quickly if a separation vector is specified, and is more suitable for situations where there are a large number of convex polygons. The core idea of ​​the GJK algorithm is to give an initial direction to two convex polygons, iteratively construct and update a simplex by constructing a Minkowski difference set, and determine whether the simplex contains the origin. If it contains the origin, the two polygons intersect, otherwise they do not intersect.

[0047] Let's take two convex polygons in two-dimensional space as an example. Figure 2The detection step diagram of the GJK algorithm shown in the figure explains the collision detection process of the GJK algorithm:

[0048] Initialization: First, an initial direction is obtained randomly. One convex polygon searches for support points in the initial direction, and the other polygon searches for support points in the opposite direction of the initial direction. In this way, two support points can be obtained.

[0049] Iterate the simplex vertices and expansion direction: Based on the two support points, calculate the Minkowski difference to get the first vertex of the simplex, and use the direction of the vertex pointing to the origin as the expansion direction for the next calculation of the support point. Calculate the second support point again through the Support function to determine whether the two simplex vertices are on both sides of the origin. If not, directly determine that the two convex polygons do not intersect. If so, select the normal vector of the current line segment facing the origin as the next expansion direction.

[0050] Determine the positional relationship between the triangle and the origin: Calculate the Minkowski difference based on the third pair of support points to get the third vertex of the simplex, and determine whether the triangle contains the origin. If so, determine that the two convex polygons intersect. If not, use the normal vectors of the two vertices closest to the origin pointing to the origin as the next expansion direction, continue to iteratively calculate the simplex vertices, and determine whether it contains the origin.

[0051] Termination condition: If the number of vertices of the envelope exceeds the dimension of the shape plus one or the number of iterations reaches the set value, the algorithm terminates.

[0052] In an embodiment of the present invention, the directed bounding box of the industrial robot and the voxel blocks within the working range are used as inputs of the GJK collision detection algorithm. The working range of the industrial robot is iteratively calculated in the sampling space using the Monte Carlo method. Through the above process, the collision situation between the current robot state and the composite CNC machining environment can be quickly obtained.

[0053] After completing the description of the collision detection method, the following describes how to expand a new node.

[0054] When expanding new nodes, it is necessary to ensure that each new node does not collide. In the face of a complex CNC machining environment, in order to improve the path search speed, in an embodiment of the present invention, an adaptive expansion step can be used to gradually expand new nodes, wherein the expansion step is adaptively updated according to the congestion level of the spatial environment.

[0055] First, the starting point is used as the current node to expand the new node for the current node; then, each time a new node is expanded, the new node expanded last time is used as the current node to expand the new node for the current node.

[0056] In one embodiment of the present invention, the method of expanding a new node for the current node may include the following contents:

[0057] Step 1020: Sampling in the spatial environment according to the current expansion step length, after obtaining the sampling node, detecting whether the sampling node will collide, if a collision occurs, adaptively updating the current expansion step length, and re-executing this step 1020 with the adaptively updated expansion step length as the current expansion step length, until it is detected that the sampling node will not collide, and then executing the next step 1022;

[0058] Step 1022: determine the sampling node as a new node expanded on the tree; determine whether the new node satisfies the direct connection condition with the end point, if not, use the new node as the new node for this expansion; otherwise, directly connect the new node with the end point to obtain a collision-free path.

[0059] In the embodiment of the present invention, if a collision of sampling nodes is detected each time sampling is performed, it indicates that the number of sampling failures increases by one, and the more sampling failures there are, the more crowded the spatial environment is, and the fewer sampling failures there are (the minimum number of sampling failures is 0, indicating that a new node is successfully obtained in the first sampling), the less crowded the spatial environment is. If the spatial environment is more crowded, the expansion step size needs to be reduced to improve the accuracy of the sampling results; if the spatial environment is very uncrowded, the expansion step size can be increased on the basis of the set expansion step size to improve the efficiency of path planning.

[0060] In the embodiment of the present invention, the current extension step size may be adaptively updated at least according to the following formula:

[0061]

[0062] Among them, k s is the adaptive expansion factor, and the expansion step after adaptive update is the product of the adaptive expansion factor and the initial expansion step. fail The number of sampling failures for the current node.

[0063] It should be noted that when the current extension step is adaptively updated according to the above formula, a first number can be set, and the extension step is updated once each time the number of sampling failures reaches the first number. The first number can be once or multiple times.

[0064] Furthermore, the extension step length can be increased according to the above formula. Specifically, a second number can be set. If the number of consecutive successful sampling reaches the second number, the extension step length can be increased according to the above formula, that is, T failSubstituting the above formula, the calculated adaptive expansion factor is a value greater than 0, and then the product of the adaptive expansion factor and the current expansion step length is used as the current increased expansion step length. For example, when the partial path for expanding a new node is node 1-node 2-node 3-node 4, when sampling node 1 to obtain node 2 is successful at one time, the product of the adaptive expansion factor and the initial expansion step length L0 is used as the expansion step length L1 for sampling node 2, and when sampling node 2 using the current expansion step length L0 to obtain node 3 is successful at one time, the result of the adaptive expansion factor and the current expansion step length L0 is used as the expansion step length L2 for sampling node 3.

[0065] Please refer to Figure 3 , is a schematic diagram of the collision-free path obtained by using the adaptive expansion step size, X obs is an obstacle, at the starting point x init To the end point x goal On the path between them, the step sizes between adjacent nodes are not exactly the same.

[0066] When the sampling node is close to the obstacle, it is more likely to collide with the industrial robot and the processing environment. In view of this situation, in one embodiment of the present invention, the idea of ​​attractive potential field and repulsive potential field is introduced to heuristically modify the position of the new node.

[0067] Specifically, after detecting that the sampling node will not collide, and before determining the sampling node as a new node expanded on the tree, the method further includes:

[0068] Step 1021: Determine whether the sampling node is located in the repulsive potential field of the obstacle; if so, determine the repulsive potential field coefficient generated by each obstacle on the sampling node, and determine the attractive potential field coefficient generated by the end point on the sampling node; update the position of the sampling node according to the repulsive potential field coefficient and the attractive potential field coefficient, and determine the updated node as a new node expanded on the tree.

[0069] Please refer to Figure 4 , is a schematic diagram of the heuristic potential field. Each obstacle corresponds to a repulsive domain formed by a corresponding action radius (circular dashed box in the figure). When the sampling node is located in the repulsive domain of the obstacle, it will be affected by the repulsive force of the obstacle and the attractive force of the end point. Under the action of the repulsive force and attractive force, a new node x will be generated. new , so that the position of the sampling node is updated to a more reasonable position, which can improve the convergence efficiency of the algorithm under the premise of reducing the possibility of collision.

[0070] Specifically, the position of the sampling node is updated according to the following formula:

[0071]

[0072] Among them, k att,prev , k rep,prev are the attractive potential field coefficient and repulsive potential field coefficient calculated at the position of the sampling node, respectively, att , k rep are the initial attractive potential field coefficient and repulsive potential field coefficient, ρ(x1,x2) is the Euclidean distance between x1 and x2, and x prev is the sampling node, x goal is the end point, x near is the node on the tree that is closest to the sampling node, v center (j) is the center point of the jth obstacle, v r (j) is the effective radius of the jth obstacle, n is the total number of obstacles, Step is the fixed expansion step, Rand() is a random number, ε is the perturbation coefficient, and x new is the updated node position.

[0073] According to the above formula, there is a situation where the repulsive force F exerted by the attractive potential field and the repulsive potential field at the original sampling node is rep and attraction F att When the sizes are the same and the directions are opposite, the update of the new node will overlap with the original sampled node (local optimum), and fall into the local optimum. To solve this phenomenon, the random perturbation ε is introduced to make the node escape from the local optimum.

[0074] In the embodiment of the present invention, after determining the new node expanded on the tree, it is determined whether the new node satisfies the direct connection condition with the end point, and the direct connection condition may be: the Euclidean distance between the new node and the end point is less than a threshold, and there is no obstacle between the new node and the end point. In this way, the new node can be directly connected to the end point, and a collision-free path is obtained.

[0075] Finally, for step 104, the collision-free path is iteratively optimized multiple times using a Gaussian sampling strategy to obtain an optimal collision-free path.

[0076] In the embodiment of the present invention, in order to ensure that the collision-free path is as short as possible, the collision-free path can be optimized by multiple iterations through a Gaussian sampling strategy. Specifically, the optimization scheme can be at least:

[0077] Step 1040: Determine a sampling range based on the collision-free path;

[0078] Step 1042: Perform Gaussian sampling that obeys the normal distribution within the sampling range to obtain Gaussian sampling points, and use the Gaussian sampling points to determine the collision-free path after this iterative optimization; use the optimized collision-free path to update the sampling range, and re-execute this step using the updated sampling range to perform the next iterative optimization, until the set number of times is reached, then terminate the update of the sampling range and obtain the optimal collision-free path.

[0079] In an embodiment of the present invention, the sampling range determined based on the collision-free path is: the path length is used as the major axis of the hyperellipsoid, and the Euclidean distance between the starting point and the ending point on the collision-free path is used as the focal length of the hyperellipsoid domain to obtain the sampling range formed by the hyperellipsoid.

[0080] Gaussian sampling that obeys normal distribution is performed within the sampling range. After obtaining the Gaussian sampling points, the Gaussian sampling points are used as new nodes expanded on the tree, and a collision-free path is re-determined using the nodes on the tree to obtain the collision-free path after this iterative optimization.

[0081] In one embodiment of the present invention, after obtaining the optimized collision-free path, the major axis C of the hyperellipsoid best will change, and the smaller value of the adaptive parameter and the feasible path length can be taken as the major axis C of the next iteration super ellipsoid best , major axis C best The update formula is as follows:

[0082]

[0083]

[0084] Among them, X elispe is a Gaussian sampling point set that follows a normal distribution, C min is the focal length of the hyperellipsoid, P Gauss is a normal probability distribution that obeys the Gaussian function, K is the optimal length of the optimized path using the parent selection mechanism of the adaptive informed potential field RRT* algorithm and C min , size(tree) is the number of nodes in the current tree, and Samples is the number of iterations initially set.

[0085] Please refer to Figure 5 , which is a schematic diagram of the sampling range formed by the hyperellipsoid. After completing the iteration according to the initial set number of iterations, an optimal collision-free path can be obtained.

[0086] This solution is tested and has high efficiency and high quality search capabilities in two-dimensional and three-dimensional complex environments. The number of iterations is set to 5000 and the initial expansion step is set to 4; please refer to Figure 6For a schematic diagram of the collision-free path search results in different complex environments of a two-dimensional scene, please refer to Figure 7 A schematic diagram of obtaining collision-free path search results in different complex environments of a three-dimensional scene.

[0087] Please refer to Figure 8 , which is a scene modeling diagram of ultrasonic strengthening robot composite CNC machining. In the path planning process of the ultrasonic strengthening robot, collision detection is required to ensure that the obtained path has no collision. The collision detection mainly includes two parts: collision detection of the ultrasonic strengthening tool head and collision detection of the robot and the CNC machine tool. That is to say, the ultrasonic strengthening tool head of the ultrasonic strengthening robot cannot collide with any workpiece and CNC machine tool parts during the process of running from the starting point to the workpiece feed point (end point). All robot joints should meet this requirement.

[0088] Next, we select two transition paths with different lengths and characteristics as objects, and select RRT*, RRT*-Connect, RRT*-Extend, Lazy-RRT*, Informed-RRT* and P-RRT* as comparison algorithms to verify the comprehensive performance of this scheme. The longest path length L after 30 independent runs is obtained. max , the average path length L mean and the shortest path length L min The result, and the maximum running time T max , average running time T mean and the shortest running time T min The result, and the maximum number of path points P max , the average number of path points P mean and the number of shortest path points P min result.

[0089] From the experimental results, we can see that among the two paths obtained by this scheme, L max , L mean and L min The mean of is the smallest, which verifies that this scheme has strong optimization ability and effective stability, which is also necessary to solve the actual path search problem and can be used to efficiently solve the collision-free path planning problem in complex scenarios. The longest running time T max , average running time T mean and the shortest running time T min The mean of is also the smallest, proving the good operating efficiency of this scheme. max , the average number of path points P mean and the number of shortest path points P min It is also the least, which reflects the excellent search efficiency of this scheme.

[0090] In the embodiments of the present invention, there are at least the following advantages:

[0091] It can perform path planning in two-dimensional and three-dimensional complex environments, and is suitable for various scenarios of collision-free path planning for industrial robots. It has strong scalability and versatility.

[0092] By using directed bounding box to build the model for the industrial robot and building the processing environment model based on the hash index voxel octree method, collision detection can be achieved efficiently and well.

[0093] The number of sampling failures is used to adaptively update the expansion step size, and the idea of ​​attractive potential field and repulsive potential field is used to heuristically modify the node position, which can effectively improve the search efficiency and reduce the possibility of collision;

[0094] Using the Gaussian sampling strategy and the parent selection mechanism to optimize the collision-free path in the dynamically formed super-ellipsoid domain can effectively improve the convergence speed of the algorithm and shorten the path length;

[0095] It can integrate path planning algorithms, offline programming, and simulation, thereby reducing the cost of manual path planning and improving production efficiency and automation.

[0096] Please refer to Fig. 9 , an embodiment of the present invention provides an ultrasonic enhanced robot obstacle avoidance path planning device, the device comprising:

[0097] A construction unit 900 is used to construct an industrial robot model and a complex CNC machining environment model;

[0098] The node expansion unit 902 is used to use the starting point and the ending point as nodes on the tree, and use the industrial robot model and the processing environment model to gradually expand new nodes on the tree from the starting point according to the adaptive expansion step length to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the congestion degree of the spatial environment;

[0099] The path optimization unit 904 is used to perform multiple iterations of optimization on the collision-free path using a Gaussian sampling strategy to obtain an optimal collision-free path.

[0100] In one embodiment of the present invention, the node expansion unit expands the new node in the following manner:

[0101] Set the new node expanded last time as the current node and execute the following command for the current node:

[0102] Sampling in the spatial environment according to the current expansion step length, after obtaining the sampling node, detect whether the sampling node will collide. If a collision occurs, the current expansion step length is adaptively updated, and the adaptively updated expansion step length is used as the current expansion step length to re-execute this step until it is detected that the sampling node will not collide, and then execute the next step;

[0103] The sampling node is determined as a new node expanded on the tree; whether the new node satisfies the direct connection condition with the end point is determined, if not, the new node is used as the new node for this expansion; otherwise, the new node is directly connected to the end point to obtain a collision-free path.

[0104] In one embodiment of the present invention, when the node extension unit adaptively updates the current extension step size, the following steps are specifically performed:

[0105] Adaptively update the current expansion step size according to the following formula:

[0106]

[0107] Among them, k s is the adaptive expansion factor, and the expansion step after adaptive update is the product of the adaptive expansion factor and the initial expansion step. fail The number of sampling failures for the current node.

[0108] In one embodiment of the present invention, after detecting that the sampling node will not collide and before determining the sampling node as a new node expanded on the tree, the node expansion unit is further used to:

[0109] Determine whether the sampling node is located in the repulsive potential field of the obstacle; if so, determine the repulsive potential field coefficient generated by each obstacle on the sampling node, and determine the attractive potential field coefficient generated by the end point on the sampling node; update the position of the sampling node according to the repulsive potential field coefficient and the attractive potential field coefficient, and determine the updated node as a new node expanded on the tree.

[0110] In one embodiment of the present invention, when the node expansion unit updates the position of the sampling node according to the repulsive potential field coefficient and the attractive potential field coefficient, the node expansion unit specifically includes:

[0111] Update the position of the sampling node according to the following formula:

[0112]

[0113] Among them, k att,prev , k rep,prev are the attractive potential field coefficient and repulsive potential field coefficient calculated at the position of the sampling node, respectively,att , k rep are the initial attractive potential field coefficient and repulsive potential field coefficient, ρ(x1,x2) is the Euclidean distance between x1 and x2, and x prev is the sampling node, x goal is the end point, x near is the node on the tree that is closest to the sampling node, v center (j) is the center point of the jth obstacle, v r (j) is the effective radius of the jth obstacle, n is the total number of obstacles, Step is the fixed expansion step, Rand() is a random number, ε is the perturbation coefficient, and x new is the updated node position.

[0114] In one embodiment of the present invention, the path optimization unit is specifically used for:

[0115] Determine the sampling range based on the collision-free path;

[0116] Gaussian sampling that obeys the normal distribution is performed within the sampling range to obtain Gaussian sampling points, and the Gaussian sampling points are used to determine the collision-free path after this iterative optimization; the sampling range is updated using the optimized collision-free path, and this step is re-executed using the updated sampling range to perform the next iterative optimization until the set number of times is reached, at which time the update of the sampling range is terminated to obtain the optimal collision-free path.

[0117] It should be noted that the ultrasonic enhanced robot obstacle avoidance path planning device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the ultrasonic enhanced robot obstacle avoidance path planning device provided in the above embodiment and the ultrasonic enhanced robot obstacle avoidance path planning method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0118] The embodiment of the present application also provides a computer device, please refer to Fig.10 The computer device includes a processor and a memory, wherein at least one instruction, at least one program, a code set or an instruction set is stored in the memory, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the ultrasonic enhanced robot obstacle avoidance path planning method provided in the above-mentioned method embodiments.

[0119] An embodiment of the present application also provides a computer-readable storage medium, on which is stored at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the ultrasonic enhanced robot obstacle avoidance path planning method provided in the above-mentioned method embodiments.

[0120] An embodiment of the present application also provides a computer program product, which includes a computer program. A processor of a computer device reads the computer program from a computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the ultrasonic enhanced robot obstacle avoidance path planning method described in any of the above embodiments.

[0121] For the convenience of description, the above system or device is described by dividing it into various modules or units according to its functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0122] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0123] Finally, it should be noted that, in this article, relational terms such as first, second, third and fourth are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0124] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An ultrasonic enhanced robot obstacle avoidance path planning method, characterized in that: include: Build industrial robot models and complex CNC machining environment models; The starting point and the end point are used as nodes on the tree, and the industrial robot model and the processing environment model are used to gradually expand new nodes on the tree according to the adaptive expansion step length starting from the starting point to obtain a collision-free path; In the process of expanding new nodes, the expansion step size is adaptively updated according to the congestion level of the spatial environment; The collision-free path is optimized multiple times by using Gaussian sampling strategy to obtain the optimal collision-free path.

2. The method according to claim 1, characterized in that Ways to expand new nodes include: Set the new node expanded last time as the current node and execute the following command for the current node: Sampling in the spatial environment according to the current expansion step length, after obtaining the sampling node, detect whether the sampling node will collide. If a collision occurs, the current expansion step length is adaptively updated, and the adaptively updated expansion step length is used as the current expansion step length to re-execute this step until it is detected that the sampling node will not collide, and then execute the next step; The sampling node is determined as a new node expanded on the tree; whether the new node satisfies the direct connection condition with the end point is determined, if not, the new node is used as the new node for this expansion; otherwise, the new node is directly connected to the end point to obtain a collision-free path.

3. The method according to claim 2, characterized in that The adaptive updating of the current expansion step size includes: Adaptively update the current expansion step size according to the following formula: Among them, k s is the adaptive expansion factor, and the expansion step after adaptive update is the product of the adaptive expansion factor and the initial expansion step. fail The number of sampling failures for the current node.

4. The method according to claim 2, characterized in that: After detecting that the sampling node will not collide, and before determining the sampling node as a new node expanded on the tree, the method further includes: Determine whether the sampling node is located in the repulsive potential field of the obstacle; If so, determine the repulsive potential field coefficient generated by each obstacle on the sampling node, and determine the attractive potential field coefficient generated by the end point on the sampling node; update the position of the sampling node according to the repulsive potential field coefficient and the attractive potential field coefficient, and determine the updated node as a new node expanded on the tree.

5. The method according to claim 4, characterized in that The updating of the position of the sampling node according to the repulsive potential field coefficient and the attractive potential field coefficient comprises: Update the position of the sampling node according to the following formula: Among them, k att,prev , k rep,prev are the attractive potential field coefficient and repulsive potential field coefficient calculated at the position of the sampling node, respectively, att , k rep are the initial attractive potential field coefficient and repulsive potential field coefficient, ρ(x1,x2) is the Euclidean distance between x1 and x2, and x prev is the sampling node, x goal is the end point, x near is the node on the tree that is closest to the sampling node, v center (j) is the center point of the jth obstacle, v r (j) is the effective radius of the jth obstacle, n is the total number of obstacles, Step is the fixed expansion step, Rand() is a random number, ε is the perturbation coefficient, and x new is the updated node position.

6. The method according to any one of claims 1 to 5, characterized in that: The Gaussian sampling strategy is used to perform multiple iterations of optimization on the collision-free path, including: Determine the sampling range based on the collision-free path; Gaussian sampling that obeys the normal distribution is performed within the sampling range to obtain Gaussian sampling points, and the Gaussian sampling points are used to determine the collision-free path after this iterative optimization; the sampling range is updated using the optimized collision-free path, and this step is re-executed using the updated sampling range to perform the next iterative optimization until the set number of times is reached, at which time the update of the sampling range is terminated to obtain the optimal collision-free path.

7. An ultrasonic enhanced robot obstacle avoidance path planning device, characterized in that: The device comprises: Construction unit, used to build industrial robot models and complex CNC machining environment models; A node expansion unit, used to use the starting point and the ending point as nodes on the tree, and to use the industrial robot model and the processing environment model to gradually expand new nodes on the tree from the starting point according to the adaptive expansion step length to obtain a collision-free path; in the process of expanding the new node, the expansion step length is adaptively updated according to the congestion degree of the spatial environment; The path optimization unit is used to perform multiple iterative optimizations on the collision-free path using a Gaussian sampling strategy to obtain an optimal collision-free path.

8. A computer device, characterized in that: The computer device includes a memory and a processor, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to implement the steps of any one of the methods described in claims 1-6.

9. 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 steps of the method described in any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The method comprises a computer program, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Citation Information

Cited By

  • High-dynamic flight system safety control method, device, equipment and storage medium

    CN119937373A

  • High-precision assembly path planning method for artificial intelligence cooperative industrial robot

    CN121821391A

  • High-precision assembly path planning method for artificial intelligence cooperating with industrial robots

    CN121821391B