Mechanical arm obstacle avoidance path planning method based on improved RRT-Connect algorithm, smearing mechanical arm and masonry robot

Abstract

The invention discloses a mechanical arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm, a smearing mechanical arm and a masonry robot. The mechanical arm obstacle avoidance path planning method comprises the steps that firstly, a fixed sampling function is adopted in a three-dimensional space formed by a starting point Xinit and a target point Xgoal of a path to generate a fixed point; and respectively and simultaneously constructing four random trees between the starting point and the fixed point and between the fixed point and the target point to carry out path search, and generating a path from the starting point to the target point. A target guiding sampling method based on a heuristic function can be introduced into the algorithm of the four-random-tree collaborative path searching mechanism, and the sampling probability is adjusted; and a better path planning effect is achieved by setting a self-adaptive step length strategy, a greedy strategy and other technical means. According to the method, the path planning of the mechanical arm is realized by setting the fixed sampling points to perform the four-way RRT-Connect algorithm, so that the rapid planning is ensured, the path time consumption is reduced, the collision with an obstacle is avoided, and the robustness of the path is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of robot path planning, and more specifically, to a robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm, a painting robot arm and a masonry robot. Background Art

[0002] In the field of construction, mud application is one of the important processes of wall construction and decoration, which traditionally relies mainly on manual operation. In particular, the coke oven masonry process is complicated, the labor efficiency is low, the refractory brick masonry workload is large and the precision requirements are high. The robot intelligent masonry of large coke oven bodies can not only save a lot of manpower, but also save energy and protect the environment during the operation process, and can also greatly shorten the construction period. With the improvement of construction efficiency and quality requirements, the automatic application technology of bricks has gradually become a mainstream demand. In this process, the use of a robotic arm for mud application can significantly improve efficiency and reduce human errors. General application robotic arms are six-degree-of-freedom robotic arms. However, the existing intelligent bionic path planning algorithm is complex to calculate, time-consuming and memory-intensive, and is difficult to apply to the path planning of high-degree-of-freedom robotic arms.

[0003] Currently, most of the relevant invention patents only propose methods for robot arm path planning, but do not take into account the complex conditions in the actual working environment. The implementation of the path planning method in special environments is not described in detail. It has not yet been applied to environmental conditions with complex and diverse obstacles, confined high-dimensional space, and dynamic changes.

[0004] The Chinese patent application with application number 202311494775.4 discloses a robot arm obstacle avoidance path planning method, which starts from solving the problems of low path accuracy and poor computational efficiency of traditional path planning methods in complex environments, focusing on improving the accuracy and optimization efficiency of path planning, and reducing the number of collision detections and computational costs. The random sampling method and collision detection strategy of traditional path planning are changed, and the pixelized planning space and optimized sampling mechanism are combined with path optimization to greatly shorten the path search and algorithm convergence time while ensuring the path quality.

[0005] The RRT-Connect (or Rapidly-exploring Random Tree) algorithm is an improved version of the RRT algorithm. The main steps of the algorithm are to use the starting point and the target point as the root node, randomly sample in the robot's workspace to find the root node closest to the sampling point in the expansion tree, and connect the nearest node with the random sampling point according to the specified step length. After collision detection, if there is no collision, the sampling point is added to the tree, and the random tree is connected by connecting the nodes to grow continuously until the sampling is stopped in the middle to complete the path search. The generation of random nodes in the RRT algorithm and the RRT-Connect algorithm is not goal-oriented, resulting in a large number of redundant points in the sampling process, and the expansion of nodes according to a fixed step length leads to a long path finding time. The turning point of the final path will cause the robot to change its acceleration greatly here, and even a singular point will appear that does not satisfy the robot kinematics. At present, some technologies (such as patent publication numbers CN119347751A and CN117029845A) have disclosed RRT path planning algorithms based on adaptive step size and / or B-spline smoothing to solve these problems. However, obstacles are complex and diverse in masonry operations, and environmental conditions change dynamically. The real-time obstacle avoidance path planning of the painting robot in this environment requires high planning speed and accuracy. At present, the problems of high computing resource consumption, slow convergence speed, and local optimal solution in path planning in complex and dynamically changing environments have not been effectively solved. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention proposes a robot obstacle avoidance path planning method, a painting robot and a masonry robot based on an improved RRT-Connect algorithm. Fixed sampling points are set to construct four random trees for a collaborative path search mechanism to realize the path planning of the robot, ensure rapid planning, reduce path time consumption, avoid collisions with obstacles, and enhance the robustness of the path.

[0007] To achieve the above objectives, in a first aspect, the present invention provides a robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm, comprising the following steps:

[0008] S1, at the starting point X of the path init (x i ,y i , z i ) and the target point X goal (x g ,y g , z g ) generates a sampling fixed point in the three-dimensional space formed by ; the generation of the fixed point includes the following steps:

[0009] Starting point X init and the target point X goal The three-dimensional midpoint Xcenter (x c ,y c , z c ) is calculated as:

[0010]

[0011] If X center If there is no collision with the obstacle, the midpoint X center As a sampling fixed point;

[0012] The midpoint X center When a collision occurs with an obstacle, a fixed sampling function is used to determine the fixed point. The steps are as follows:

[0013] (1) Constructing the sphere S: When X center When in the obstacle area, the initial radius R of the sphere S init for: ;in, is the scaling factor, which is adjusted according to the complexity of the scene and the distribution of obstacles; The starting point X init With the target point X goal Euclidean distance; construct center is the center of the circle, R init is a sphere of radius S;

[0014] (2) Define plane L: Randomly select a point P(x p ,y p , z p ), make sure not to init and X goal The plane L is connected by X center , point P and direction vector Determine together, the plane equation is:

[0015]

[0016] (3) Construct a path through X center And the straight line l is perpendicular to the plane L; its direction vector is the normal vector of the plane L, and the equation form is:

[0017]

[0018] Where A, B, C are the normal vector components of plane L:

[0019]

[0020] Find the intersection points X1 and X2 of the straight line l and the sphere S;

[0021] (4) Intersection point screening, check whether X1 and X2 are in free space: If neither X1 nor X2 collides with obstacles, select the first feasible intersection point as X1. fixed ; If both collide with obstacles, increase the radius R init Repeat the above steps until a feasible X is found. fixed ;

[0022] S2. Construct four random trees simultaneously between the starting point and the fixed point, and between the fixed point and the target point to perform path search, and generate a path from the starting point to the target point.

[0023] By adopting the above technical solution, the present invention introduces fixed sampling points, decomposes the path search into four random trees collaborative path search, and simultaneously improves the search efficiency, path quality and algorithm stability, adapting to complex environments and optimization requirements. The algorithm for generating fixed points from the fixed sampling function proposed in the present invention can evenly distribute key sampling points in the search space, guide the expansion direction of the tree, and improve the coverage and connectivity of the path search. Compared with traditional random sampling, fixed sampling points reduce invalid expansion, improve search efficiency, and ensure the stability and feasibility of the path in complex environments. In addition, fixed sampling combined with the four random trees collaborative strategy reduces path redundancy, improves the global optimality of the search, and makes the generated path smoother and meets the optimization requirements.

[0024] Furthermore, the collaborative path search algorithm of the four random trees introduces a target-guided sampling method based on a heuristic function, adjusts the sampling probability, makes the sampling process more directional, and reduces the generation of invalid nodes.

[0025] Furthermore, the heuristic function h(x) evaluates the distance from the current node to the target node, and calculates the probability P of the target node as a sampling point based on the distance. goal ;

[0026]

[0027] in, is the coordinate of the current sampling node; is the coordinate of the target node;

[0028] Define an increasing target point sampling probability P goal , the probability increases as h(x) decreases; when the current node is close to the target area, the probability of the target point being sampled is higher; use the Sigmoid type increasing function to describe this probability relationship:

[0029]

[0030] Among them, α is the rate at which the control probability increases. The larger the value, the faster the probability of the target node as a sampling point changes; dmax The maximum distance of the preset sampling area is used for normalized distance;

[0031] Calculate the selection probability of random points and target points in the sampling area. In each sampling, according to the calculated P goal Decide whether to select the target point or randomly generate a point in the sampling area: Generate a random number r∈[0,1]; if r <P goal , then select the target point X goal As a sampling point; otherwise, randomly select a sampling point X in the entire space rand , and perform iterative calculations.

[0032] Furthermore, the collaborative path search algorithm of the four random trees sets an adaptive step size strategy, dynamically adjusts the step size according to the complexity of the environment, the target distance and the density of the tree, and adds a greedy strategy in the node expansion to further improve the efficiency of random tree expansion; the adaptive initialization step size formula is:

[0033]

[0034] Among them, M size is the environmental size, Express Round up; w1 is the environmental complexity coefficient, which is based on the area, distance and number of obstacles. It comprehensively calculates the ratio of the obstacle area to the overall area, the average of the distances of adjacent obstacles, the inverse of the number of obstacles, and multiplies them by different weight coefficients; in the process of expanding to the target point, the collected environmental information is used. When encountering narrow areas or areas with many obstacles, the RRT-Connect algorithm is improved to adaptively adjust the step size:

[0035]

[0036] Among them, w2 is the step size reduction coefficient until the expansion tree covers the current area, and then the initial step size is restored;

[0037] When w1 is small, the default node is in the open area. rand =X goa1 , the step size is dynamic, and it keeps growing along the direction of the target point until it encounters an obstacle.

[0038] Furthermore, after the collaborative path search of the four random trees is completed, a basic pruning operation is performed, and a path pruning strategy based on a triangle inequality is used to optimize the path in an iterative process to remove redundant nodes in the random trees to the greatest extent.

[0039] Furthermore, after the path is optimized, a cubic B-spline curve is used for smoothing to generate a path from the starting point to the target point, so as to obtain a path with the minimum cost and no collision from the starting point to the target point.

[0040] In a second aspect, the present invention provides a robotic arm that uses the robotic arm obstacle avoidance path planning method as described above to achieve path planning between its original position and target position.

[0041] In a third aspect, the present invention provides a coating robot arm for automatically coating mud on bricks, and adopts the robot arm obstacle avoidance path planning method as described above to realize path planning between its original position and the brick position.

[0042] In a final aspect, the present invention provides a masonry robot, comprising:

[0043] The truss and gripper unit moves and fixes the bricks to be smeared with mud within the working area of ​​the smearing robot arm;

[0044] The smearing robot arm as described above is loaded with a slurry pipe and automatically applies slurry to the bricks.

[0045] Compared with the prior art, the present invention has the following technical effects:

[0046] (1) The present invention introduces fixed sampling points and decomposes the path search into four random trees collaborative path search, which simultaneously improves the search efficiency, path quality and algorithm stability, and adapts to complex environments and optimization requirements. The algorithm for generating fixed points using the fixed sampling function introduced can evenly distribute key sampling points in the search space, guide the expansion direction of the tree, and improve the coverage and connectivity of the path search. Compared with traditional random sampling, fixed sampling points reduce invalid expansion, improve search efficiency, and ensure the stability and feasibility of the path in complex environments. In addition, fixed sampling combined with the four random trees collaborative strategy reduces path redundancy and improves the global optimality of the search, making the generated path smoother and in line with optimization requirements.

[0047] (2) The present invention also introduces a target-guided sampling method based on a heuristic function to adjust the sampling probability, making the sampling process more directional and reducing the generation of invalid nodes.

[0048] (3) The present invention improves the efficiency of random tree expansion by setting an adaptive step size strategy and a greedy strategy; after basic pruning operations, a path pruning strategy based on triangle inequality is used to remove redundant nodes in the random tree to the greatest extent; and a cubic B-spline curve is used to smooth the path to obtain a path with the minimum cost and no collision from the starting point to the target point. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention and its features and advantages will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following accompanying drawings.

[0050] Figure 1 is a schematic diagram of a masonry robot provided by an embodiment of the present invention;

[0051] Figure 2 is a flow chart of a robot arm obstacle avoidance path planning algorithm provided by an embodiment of the present invention;

[0052] Figure 3 is a schematic diagram of the RRT-Connect principle for introducing fixed sampling points provided by an embodiment of the present invention;

[0053] Figure 4 It is a schematic diagram of the principle of a fixed sampling function in a three-dimensional space provided by an embodiment of the present invention;

[0054] Figure 5 Schematic diagram of the principle of adaptive step size provided by an embodiment of the present invention;

[0055] Figure 6 is a schematic diagram of the path pruning principle provided by an embodiment of the present invention;

[0056] Figure 7 A schematic diagram of a B-spline curve provided by an embodiment of the present invention;

[0057] Figure 8 This is the effect diagram of the traditional RRT-Connect algorithm path planning;

[0058] Fig. 9 This is a rendering of a path planning algorithm according to an embodiment of the present invention;

[0059] In the figure, 1. Truss; 2. Painting robot arm; 3. Brick clamping mechanism. DETAILED DESCRIPTION

[0060] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0061] It should be noted that the device structure and drawings of the present invention mainly describe the principles of the present invention. In terms of the technology of the design principle, the settings of the power mechanism, power supply system and control system of the device are not fully described. On the premise that the technical personnel in this field understand the principles of the above invention, the details of its power mechanism, power supply system and control system can be clearly known. The control method can be automatic control through a controller. The control algorithm of the controller can be realized by simple programming by technical personnel in this field, and the undetailed technology involved in the following effect embodiments is a searchable prior art.

[0062] In addition, the order of execution of actions, steps, etc. in the devices and methods shown in the claims, specifications and drawings can be implemented in any order as long as there is no special explicit limitation on the order and the output of the previous processing is not used in the subsequent processing.

[0063] Example

[0064] See also Figure 1 In this embodiment, the coating robot arm 2 in the masonry robot applies mud to the bricks pulled by the brick clamping mechanism 3 on the truss 1 as an example. The coating robot arm 2 realizes obstacle avoidance path planning between its original position and the brick position through the following technical means.

[0065] See also Figure 2 and Figure 3 This embodiment provides a robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm, comprising the following steps:

[0066] S1, at the starting point X of the path init and the target point X goal Generate a fixed point in the three-dimensional space formed by Figure 4 , generating fixed points includes the following steps:

[0067] Starting point X init (x i ,y i , z i ) and the target point X goal (x g ,y g , z g ) of the three-dimensional midpoint X center (x c ,y c , z c ) is calculated as:

[0068]

[0069] If X center If there is no collision with the obstacle, the midpoint X center as a fixed sampling point.

[0070] The midpoint X center When colliding with an obstacle, a fixed sampling function is used to determine the fixed point. The steps are as follows:

[0071] (1) Sphere and initial radius

[0072] When X center When in the obstacle area, the initial radius R init for: ;in, is a scaling factor, which is adjusted according to the complexity of the specific scene and the distribution of obstacles. When there are fewer obstacles in the environment and the space is relatively open, a larger α value can be selected to speed up the search. When there are a large number of obstacles or narrow passages in the environment, a smaller α value should be selected to improve local obstacle avoidance capabilities. In an open environment, a larger α is selected. A large radius directly expands the search range and reduces the number of iterations to speed up the global search; in a complex environment, a smaller α is selected to avoid collisions through a refined search with a small radius. Although the single search range is small, the overall search efficiency is higher. The starting point X init With the target point X goal The Euclidean distance of center is the center of the circle, R init The equation of the sphere S in three-dimensional space with radius is:

[0073]

[0074] (2) Plane definition

[0075] Randomly select a point P(x p ,y p , z p ), make sure not to init and X goal The plane L consists of X center , point P and direction vector Determine together, the plane equation is

[0076]

[0077] (3) Intersection point of the vertical line and the sphere

[0078] Construct a path through X center And it is perpendicular to the straight line l of plane L. Its direction vector is the normal vector of plane L, and its equation form is:

[0079]

[0080] Where A, B, C are the normal vector components of plane L, then

[0081]

[0082] Find the intersection points X1 and X2 of the straight line l and the sphere S.

[0083] (4) Intersection point screening

[0084] Check if X1 and X2 are in free space: If neither X1 nor X2 has any collision with any obstacle, select the first feasible intersection point as X1. fixed; If both collide with obstacles, increase the radius R init Repeat the above steps until a feasible X is found. fixed .

[0085] S2. Construct four random trees simultaneously between the starting point and the fixed point, and between the fixed point and the target point to perform a bidirectional path search, and generate a path from the starting point to the target point.

[0086] See also Figure 3 , at the starting point X init 、Target point X goal and fixed point X fixed At the same time, four random trees T1, T2, T3 and T4 are established, and the X rand The nearest node is X nearest , if X nearest With X rand If there is no collision between init Expand the new node X with a certain step size δ new , the new node X in the first tree new After the expansion is completed, the second tree T2 will be X fixed As the root node, with the new target point X new As the expansion direction. At the same time, the second tree expands in a slightly different way. First, it will expand the first step to get X' new , if there is no collision, continue to expand step 2 in the same direction until the expansion fails or X′ new =X new Indicates that it is connected to the first tree. Similarly, select T4 from X''' rand The nearest node is X′′′ nearest , if X′′′ nearest With X''' rand If there is no collision between goal Expand the new node X''' with a certain step size δ' new , in the new node X''' in the 4th tree new After the expansion is completed, the third tree T3 will be X fixed As the root node, take the new target point X′′′ new As the expansion direction. At the same time, the fourth tree is expanded in a slightly different way. First, it will expand the first step to get X'' new , if there is no collision, continue to expand step 2 in the same direction until the expansion fails or X′′ new =X′′′ new It indicates that it is connected to the third tree and the whole algorithm ends.

[0087] As a specific implementation example, the algorithm of the four random tree collaborative path search mechanism in step S2 introduces a heuristic function to dynamically adjust the sampling probability, making the sampling process more directional, thereby reducing the generation of invalid nodes and improving the sampling efficiency and the convergence of path planning.

[0088] Preferably, the heuristic function h(x) evaluates the distance from the current node to the target node, and calculates the probability P of the target node as a sampling point based on the distance. goal ;

[0089]

[0090] in, is the coordinate of the current sampling node; is the coordinate of the target node;

[0091] Define an increasing target point sampling probability P goal , the probability increases as h(x) decreases; when the current node is close to the target area, the probability of the target point being sampled is higher; use the Sigmoid type increasing function to describe this probability relationship:

[0092]

[0093] Among them, α is the rate at which the control probability increases. The larger the value, the faster the probability of the target node as a sampling point changes; d max The maximum distance of the preset sampling area is used for normalized distance;

[0094] Calculate the selection probability of random points and target points in the sampling area. In each sampling, according to the calculated P goal Decide whether to select the target point or randomly generate a point in the sampling area: Generate a random number r∈[0,1]; if r <P goal , then select the target point X goal As a sampling point; otherwise, randomly select a sampling point X in the entire space rand , and perform iterative calculations.

[0095] As a specific implementation example, the algorithm of the four random tree collaborative path search mechanism in step S2 sets an adaptive step size strategy, dynamically adjusts the step size according to the complexity of the environment, the target distance and the density of the tree, and adds a greedy strategy in the node expansion to further improve the efficiency of random tree expansion; see Figure 5 , the adaptive initialization step size formula is:

[0096]

[0097] Among them, M size is the environmental size, Express Round up; w1 is the environmental complexity coefficient, which is based on the area, distance and number of obstacles. It comprehensively calculates the ratio of the obstacle area to the overall area, the average of the distances of adjacent obstacles, the inverse of the number of obstacles, and multiplies them by different weight coefficients; in the process of expanding to the target point, the collected environmental information is used. When encountering narrow areas or areas with many obstacles, the RRT-Connect algorithm is improved to adaptively adjust the step size:

[0098]

[0099] Among them, w2 is the step size reduction coefficient until the expansion tree covers the current area, and then the initial step size is restored;

[0100] When w1 is small, the default node is in the open area. rand =X goa1 , the step size is dynamic, and it keeps growing along the direction of the target point until it encounters an obstacle.

[0101] After the four random tree collaborative path search mechanism is completed, as a further process, a series of collision-free path point sets Ω from the initial point to the target point are obtained. First, the path point I is connected from the initial point. If there is no collision between the initial point and the path point I, the initial point is then connected to the path point II, and so on, until a collision occurs between the two points. Then, the previous point of the path point where the collision occurs is retained, and the above operation is performed again with the node as the parent node until the target point is reached.

[0102] Simple pruning: Start from the target point and connect the previous node, repeating the above process until you reach the starting point. Figure 6 (a) in the figure compares the lengths of the paths after removing redundant nodes from the initial and target points, retaining the nodes of the shorter path, and connecting the retained path nodes.

[0103] After obtaining the above path, pruning is performed again to create smaller triangles tangent to the obstacles and effectively interpolate between the path points. The pruning process is as follows: Figure 6 As shown in (b), after the initial path is generated, the intermediate paths between all nodes are assumed to not collide with obstacles. i+2 Start at Z i+1 and Z i+2 Linear interpolation is performed between Figure 6 As shown in (b), until Z i The nodes after interpolation can be connected without collision. At this time, a temporary node M will be generated. After M is found, i and M until the interpolation point and Z iThe connection does not collide with obstacles, and a new node N is generated, which will replace node Z. i+1 , the update path is Z i →N→Z i+2 The above process is repeated between all nodes. Through the pruning process, the path is pushed to the obstacle, effectively reducing the path length.

[0104] Based on the optimized path, a cubic B-spline curve is used for smoothing. The specific work is as follows:

[0105] The cubic B-spline curve is used to smooth the obstacle avoidance path. The equation of the k-order B-spline curve is:

[0106]

[0107] in, Represents the control point, which determines the control polygon of the B-spline; N i,k (u) represents the k-order B-spline basis function. And the recursive formula of the k-order B-spline basis function is:

[0108]

[0109]

[0110] Wherein, k represents the order of the B-spline function; i represents the sequence number of the B-spline function.

[0111] Cubic B-spline curve segment:

[0112]

[0113] When a control point d is given i When , the above formula can be used to find a series of curve points that satisfy the cubic B-spline.

[0114] like Figure 7 In the figure, d1d'1 and d2d'2 are the midlines of triangles d0d1d2 and d1d2d3 respectively. The starting point of the B-spline is selected as 1 / 3 of d1d'1, with a direction horizontal to the base d0d2 and a length of 1 / 2. The end point is selected as 1 / 3 of d2d'2, with a direction horizontal to the base d1d3 and a length of 1 / 2. The adjustment of each control point will only affect the local curvature characteristics of the curve, but will not affect the overall performance of the entire curve. By merging these local segmented curves, a continuous B-spline curve can be spliced.

[0115] See also Figure 8 and Fig. 9 For the same path planning requirements (Figure a is a two-dimensional path planning, Figure b is a three-dimensional path planning), the path finally planned by the traditional RRT-Connect algorithm has many redundant points ( Figure 8 (a) and (b)), the redundant points of the path implemented in this embodiment are significantly reduced ( Fig. 9 (a1) and (b1) in Figure 2), after pruning and smoothing, the path becomes clearer ( Fig. 9 For two-dimensional path planning, the path planning of the traditional RRT-Connect algorithm takes an average of 1.025 seconds, while the present embodiment takes an average of 0.5 seconds after pruning and smoothing. For three-dimensional path planning, the path planning of the traditional RRT-Connect algorithm takes an average of 1.85 seconds, while the present embodiment takes an average of 0.82 seconds after pruning and smoothing.

[0116] From the above embodiments and their effects, it can be found that the robot arm obstacle avoidance path planning method based on the improved RRT-Connect algorithm of the present invention implements the path planning of the robot arm by setting fixed sampling points for a four-random tree collaborative search mechanism, which can achieve fast planning, reduce path time consumption, avoid collisions with obstacles, and enhance the robustness of the path.

[0117] It can be understood that the above-mentioned robot arm obstacle avoidance path planning method is not only applicable to painting robot arms, but also applicable to other types of robot arms to realize point-to-point obstacle avoidance path planning.

[0118] In summary, the present invention provides a robot arm obstacle avoidance path planning method based on the improved RRT-Connect algorithm, a painting robot arm and a masonry robot. First, at the starting point X of the path, init and the target point X goal A fixed point is generated in the constructed three-dimensional space using a fixed sampling function; four random trees are simultaneously constructed between the starting point and the fixed point, and between the fixed point and the target point to search and generate a path from the starting point to the target point. The algorithm of the four random tree collaborative path search mechanism can also introduce a target-guided sampling method based on a heuristic function to adjust the sampling probability; a better path planning effect is achieved by setting technical means such as an adaptive step size strategy and a greedy strategy. The present invention realizes the path planning of the robotic arm by setting a fixed sampling point to perform a four-random tree collaborative search mechanism, ensuring rapid planning, reducing path time consumption, avoiding collisions with obstacles, and enhancing the robustness of the path.

[0119] Those skilled in the art should understand that those skilled in the art can implement variations by combining the prior art and the above embodiments, which will not be described in detail here. Such variations do not affect the essential content of the present invention, and will not be described in detail here.

[0120] The above describes the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the above-mentioned specific embodiments, and the systems and structures that are not described in detail should be understood to be implemented in a common manner in the art; any technician familiar with the art can use the above-disclosed methods and technical contents to make many possible changes and modifications to the technical solutions of the present invention without departing from the scope of the technical solutions of the present invention, or modify them into equivalent embodiments of equivalent changes, which does not affect the essential content of the present invention. Therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention are still within the scope of protection of the technical solutions of the present invention.

Claims

1. A robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm, characterized in that: The following steps are involved: S1, at the starting point X of the path init (x i ,y i , z i ) and the target point X goal (x g ,y g , z g ) generates a sampling fixed point in the three-dimensional space formed by ; the generation of the sampling fixed point includes the following steps: Starting point X init and the target point X goal The three-dimensional midpoint X center (x c ,y c , z c ) is calculated as: If X center If there is no collision with the obstacle, the midpoint X center As a sampling fixed point; The midpoint X center When a collision occurs with an obstacle, a fixed sampling function is used to determine the fixed point. The steps are as follows: (1) Constructing the sphere S: When X center When in the obstacle area, the initial radius R of the sphere S init for: ;in, is the scaling factor, which is adjusted according to the complexity of the scene and the distribution of obstacles; The starting point X init With the target point X goal Euclidean distance; construct center is the center of the circle, R init is a sphere of radius S; (2) Define plane L: Randomly select a point P(x p ,y p , z p ), make sure not to init and X goal The plane L is connected by X center , point P and direction vector Determine together, the plane equation is: (3) Construct a path through X center And the straight line l is perpendicular to the plane L; its direction vector is the normal vector of the plane L, and the equation form is: Where A, B, C are the normal vector components of plane L: Find the intersection points X1 and X2 of the straight line l and the sphere S; (4) Intersection point screening, check whether X1 and X2 are in free space: If neither X1 nor X2 collides with obstacles, select the first feasible intersection point as X1. fixed ; If both collide with obstacles, increase the radius R init Repeat the above steps until a feasible X is found. fixed ; S2. Construct four random trees simultaneously between the starting point and the fixed point, and between the fixed point and the target point to perform path search, and generate a path from the starting point to the target point.

2. According to claim 1, a robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm is characterized in that: The collaborative path search algorithm of the four random trees introduces a target-guided sampling method based on a heuristic function to adjust the sampling probability.

3. According to claim 2, a robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm is characterized in that: The heuristic function h(x) evaluates the distance from the current node to the target node and calculates the probability P of the target node as a sampling point based on the distance. goal ; in, is the coordinate of the current sampling node; is the coordinate of the target node; Define an increasing target point sampling probability P goal , the probability increases as h(x) decreases; when the current node is close to the target area, the probability of the target point being sampled is higher; use the Sigmoid type increasing function to describe this probability relationship: Among them, α is the rate at which the control probability increases. The larger the value, the faster the probability of the target node as a sampling point changes; d max The maximum distance of the preset sampling area is used for normalized distance; Calculate the selection probability of random points and target points in the sampling area. In each sampling, according to the calculated P goal Decide whether to select the target point or randomly generate a point in the sampling area: Generate a random number r∈[0,1]; if r <P goal , then select the target point X goal As a sampling point; otherwise, randomly select a sampling point X in the entire space rand , and perform iterative calculations.

4. A robot arm obstacle avoidance path planning method based on an improved RRT-Connect algorithm according to claim 1 or 2, characterized in that: The collaborative path search algorithm of the four random trees sets an adaptive step size strategy, dynamically adjusts the step size according to the complexity of the environment, the target distance and the density of the tree, and adds a greedy strategy in node expansion; The adaptive initialization step size formula is: Among them, M size is the environmental size, Express Round up; w1 is the environmental complexity coefficient, which is based on the area, distance and number of obstacles. It comprehensively calculates the ratio of the obstacle area to the overall area, the average of the distances of adjacent obstacles, the inverse of the number of obstacles, and multiplies them by different weight coefficients; in the process of expanding to the target point, the collected environmental information is used. When encountering narrow areas or areas with many obstacles, the RRT-Connect algorithm is improved to adaptively adjust the step size: Among them, w2 is the step size reduction coefficient until the expansion tree covers the current area, and then the initial step size is restored; When w1 is small, the default node is in the open area. rand =X goa1 , the step size is dynamic, and it keeps growing along the direction of the target point until it encounters an obstacle.

5. The method for obstacle avoidance path planning of a manipulator based on an improved RRT-Connect algorithm according to claim 1 or 2, characterized in that: After the four random tree path search is completed, a basic pruning operation is performed, and a path pruning strategy based on triangle inequality is used to optimize the path in the iterative process.

6. The method for obstacle avoidance path planning of a manipulator based on an improved RRT-Connect algorithm according to claim 5, characterized in that: After the path is optimized, it is smoothed using a cubic B-spline curve to generate a path from the starting point to the target point.

7. A robotic arm, characterized in that: The robot arm obstacle avoidance path planning method as described in any one of claims 1 to 6 is used to realize the path planning between its original position and the target position.

8. A coating robot arm for automatically coating bricks with mud, characterized in that: The robot arm obstacle avoidance path planning method as described in any one of claims 1 to 6 is used to realize the path planning between its original position and the brick position.

9. A masonry robot, characterized in that: include: The truss and gripper unit moves and fixes the bricks to be smeared with mud within the working area of ​​the smearing robot arm; The smearing robot arm as described in claim 8 is loaded with a mud pipe and automatically applies mud to bricks.

Images