Robot path planning method based on improved RRT-Connect algorithm
The improved RRT-Connect algorithm addresses inefficiencies and path smoothness issues by employing dynamic sampling, adaptive step size, and B-spline curve construction, resulting in efficient and smooth robot paths.
Patent Information
- Application Number
- CN202411098301.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-12-03
AI Technical Summary
Existing RRT-Connect algorithms for robot path planning suffer from low search efficiency, high path cost, and non-smooth paths, failing to meet real-time requirements in complex environments.
An improved RRT-Connect algorithm with dynamic restricted sampling, target-biased adaptive step size, and greedy node trimming, combined with quadratic B-spline curve construction to smooth paths.
Enhances search efficiency, reduces path cost, and ensures smooth, kinematically feasible paths for mobile robots in complex environments.
Smart Images

Figure QLYQS_1 
Figure QLYQS_2 
Figure QLYQS_4
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of path planning, and in particular to a robot path planning method based on an improved RRT-Connect algorithm. Background Art
[0002] With the continuous breakthroughs in related high-tech technologies such as artificial intelligence, integrated chips, computer vision, satellite navigation and 5G communication technology, the field of robotics has also achieved rapid development. Robots are widely used in industry, agriculture, military, commerce, service, medical and other industries; mobile robot path planning is an important issue that needs to be urgently solved in the development of the robotics industry. Whether the robot can achieve autonomous obstacle avoidance and select the optimal path is directly related to the safety and economy of the mobile robot's work; in complex environments, mobile robots may make mistakes, causing damage to the robot and buildings, and even injuries to people, resulting in mission failure and huge economic losses; the path cost and planning efficiency of the mobile robot during its movement are important evaluation indicators of the mobile robot's work quality, so how to efficiently plan the path is crucial to the development of the robot industry. Research on mobile robots is of great significance. According to different known conditions of the environment, path planning can be divided into two categories: global path planning and local path planning. This chapter mainly studies the global path planning of mobile robots. In recent years, there have been many research advances in the global path planning of mobile robots. For example, the Informed-RRT algorithm, RRT-Connect and RRT*-Connect algorithms. However, the Informed-RRT algorithm has problems such as low search efficiency and high path cost. Although the planning efficiency of the RRT-Connect algorithm is significantly improved, it has problems such as low sampling efficiency, large expansion randomness, and uneven planning path. Although the RRT*-Connect algorithm can obtain an asymptotic optimal solution, the search time cost is too high and it is difficult to meet real-time requirements. Summary of the invention
[0003] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a robot path planning method based on an improved RRT-Connect algorithm to solve the problems of low search efficiency and non-smooth path of the RRT-Connect algorithm;
[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:
[0005] The robot path planning method based on the improved RRT-Connect algorithm is specifically divided into the following steps:
[0006] Step 1: Initialize parameter information and build a map, set parameters such as the maximum number of iterations, starting point and target point location;
[0007] Step 2: From the starting point q start and the target point q goal Start to build two random trees T1 and T2, sample in the dynamically restricted sampling area, and get the sampling point q rand ;
[0008] Step 3: Select the sampling point q in T1 rand The nearest node q nearest1 ;
[0009] Step 4: Adopt the target biased adaptive step size strategy, from q nearest1 With a fixed step size S, rand After expansion, the adaptive step size S * Towards the target point q goal Expand to get new node q new1 ;
[0010] Step 5: Detect q nearest1 With q new1 Whether the line between them collides with an obstacle, if there is no collision, the new node q new1 Add random tree T1, otherwise resample and expand with random tree T2;
[0011] Step 6: After T1 is successfully expanded, T2 will expand the new node q according to the above steps. new2 After the collision detection is passed, the new node q new2 Add random tree T2, otherwise resample and expand with random tree T1 instead;
[0012] Step 7: Connect the two random trees to obtain an initial path consisting of broken line segments;
[0013] Step 8: Use the greedy strategy to trim redundant nodes, shorten the path length, construct a quadratic B-spline curve between adjacent nodes to smooth the path, and solve the kinematic constraint problem during the robot's movement;
[0014] Furthermore, the dynamically restricted sampling area Range in step 2 satisfies the following formula:
[0015]
[0016] Where N is the number of random tree expansions, K1 and K2 are algorithm thresholds;
[0017] Furthermore, the target in step 4 is biased towards the adaptive step size strategy S * The calculation formula is as follows:
[0018]
[0019] Where G(x) is the target bias function, and its value determines the target bias degree of node expansion. μ is the initial value of the function, and φ is the function change rate coefficient. The target bias function G(x) has the following independent variable x:
[0020]
[0021] Where Δx is the increment of the independent variable, check() is the collision detection function, and the return value is 1 if the random tree collides with the obstacle, otherwise the return value is 0;
[0022] Furthermore, the greedy strategy in step 8 prunes redundant nodes and constructs a quadratic B-spline curve to smooth the path as follows:
[0023] First, the starting point q start and the target point q goal Directly connected, if there is no collision with obstacles, the line connecting the starting point and the target point is directly used as the path. If a collision occurs, the previous node of the end point is connected to the starting point q start Connected, follow the above steps, perform collision detection on the line connecting each node and the starting point in turn, until no collision occurs, and record the node as the new path node q new1 , and then the new path node q new1 As the new starting point, repeat the above process until a point that matches the target point q is found. goal The process ends when a new path node is directly connected without collision. start 、Target point q goal Connecting them in sequence will get the optimized path after clipping. The equation of the quadratic B-spline curve is as follows:
[0024]
[0025] where F i,2 (t) is the quadratic B-spline basis function, P i is the control point;
[0026] The technical solution provided by the present invention has the following beneficial effects:
[0027] 1. The present invention sets a dynamic restricted sampling area, which limits the sampling range of the algorithm, reduces the randomness of the sampling process and ensures the probability completeness of the algorithm, effectively improving the algorithm search efficiency;
[0028] 2. The present invention adopts a target-biased adaptive step-size strategy, which improves the target orientation in the random tree expansion process and avoids the algorithm from falling into a local trap;
[0029] 3. The present invention uses a greedy strategy to trim redundant nodes, shorten the path length, and construct a quadratic B-spline curve between adjacent nodes to smooth the path, thereby obtaining a smooth path and solving the kinematic constraint problem during robot movement. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] The present invention will be further described below in conjunction with the accompanying drawings, which are as follows:
[0031] Figure 1 A flowchart of a robot path planning method based on an improved RRT-Connect algorithm of the present invention
[0032] Figure 2 Schematic diagram of dynamically limiting the sampling area
[0033] Figure 3 Schematic diagram of target bias adaptive step size
[0034] Figure 4 Clipping diagram for path DETAILED DESCRIPTION
[0035] In order to make the implementation objectives, technical means and advantages achieved by the present invention easier to understand, the present invention is further described below with reference to specific illustrations.
[0036] The present invention proposes a robot path planning method based on the improved RRT-Connect algorithm, sets a dynamic restricted sampling area, limits the algorithm sampling range, adopts a target biased adaptive step size strategy to guide the random tree expansion, uses a greedy strategy to trim redundant nodes, and constructs a quadratic B-spline curve between adjacent nodes to smooth the path, thereby solving the path planning problem of the mobile robot. Figure 1 As shown, the path planning method includes the following steps:
[0037] Step 1: Initialize parameter information and build a map, set parameters such as the maximum number of iterations, starting point and target point location;
[0038] Step 2: From the starting point q start and the target point q goal Start by building two random trees T1 and T2, sampling in the dynamically restricted sampling area, such as Figure 2 As shown, the sampling point q is obtained rand ;
[0039] The dynamic limited sampling area Range satisfies the following formula:
[0040]
[0041] Where N is the number of random tree expansions, K1 and K2 are algorithm thresholds;
[0042] Step 3: Select the sampling point q in T1 rand The nearest node q nearest1 ;
[0043] Step 4: Adopt the target biased adaptive step size strategy, from q nearest1 With a fixed step size S, rand After expansion, the adaptive step size S * Towards the target point q goal Expand to get new node q new1 ,like Figure 3 As shown;
[0044] Adaptive step size S * The calculation formula is as follows:
[0045]
[0046] Where G(x) is the target bias function, and its value determines the target bias degree of node expansion. μ is the initial value of the function. is the function change rate coefficient, and the target bias function G(x) independent variable x satisfies the following formula:
[0047]
[0048] Where Δx is the increment of the independent variable, check() is the collision detection function, and the return value is 1 if the random tree collides with the obstacle, otherwise the return value is 0;
[0049] Step 5: Detect q nearest1 With q new1 Whether the line between them collides with an obstacle, if there is no collision, the new node q new1 Add random tree T1, otherwise resample and expand with random tree T2 instead;
[0050] Step 6: After T1 is successfully expanded, T2 will expand the new node q according to the above steps. new2 After the collision detection is passed, the new node q new2 Add random tree T2, otherwise resample and expand with random tree T1 instead;
[0051] Step 7: Connect the two random trees to obtain an initial path consisting of broken line segments;
[0052] Step 8: Use the greedy strategy to trim redundant nodes, shorten the path length, construct a quadratic B-spline curve between adjacent nodes to smooth the path, and solve the kinematic constraint problem during the robot's movement;
[0053] The process of using the greedy strategy to prune redundant nodes and construct a quadratic B-spline curve to smooth the path is as follows:
[0054] First, the starting point q start and the target point q goal Directly connected, if there is no collision with obstacles, the line connecting the starting point and the target point is directly used as the path. If a collision occurs, the previous node of the end point is connected to the starting point q start Connected, follow the above steps, perform collision detection on the line connecting each node and the starting point in turn, until no collision occurs, and record the node as the new path node q new1 , and then the new path node q new1 As the new starting point, repeat the above process until a point that matches the target point q is found. goal The process ends when a new path node is directly connected without collision. start 、Target point q goal Connect them in sequence to get the optimized path after cutting, such as Figure 4 As shown, the quadratic B-spline curve equation is as follows:
[0055] The quadratic B-spline curve equation is as follows:
[0056]
[0057] where F i,2 (t) is the quadratic B-spline basis function, P i is the control point;
[0058] After the above steps, a robot travel path with low path cost and smoothness is finally generated.
[0059] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A robot path planning method based on the improved RRT-Connect algorithm, It is characterized in that The method is specifically divided into the following steps: Step 1: Initialize parameter information and build a map, set parameters such as the maximum number of iterations, starting point and target point location; Step 2: Step 2: From the starting point q start and the target point q goal Start building two random trees T 1 and T 2 , sampling is performed in the dynamic restricted sampling area to obtain the sampling point q rand ; Step 3: Select T 1 Medium distance sampling point q rand The nearest node q nearest1 ; Step 4: Adopt the target biased adaptive step size strategy, from q nearest1 With a fixed step size S, rand After expansion, the adaptive step size S * Towards the target point q goal Expand to get new node q new1 ; Step 5: Detect q nearest1 With q new1 Whether the line between them collides with an obstacle, if there is no collision, the new node q new1 Add random tree T 1 , otherwise resample and replace with random tree T 2 To expand; Step 6: T 1 After the expansion is successful, it will be replaced by T 2 Follow the same steps to expand the new node q new2 After the collision detection is passed, the new node q new2 Add random tree T 2 , otherwise resample and replace with random tree T 1 To expand; Step 7: Connect the two random trees to obtain an initial path consisting of broken line segments; Step 8: Use the greedy strategy to trim redundant nodes, shorten the path length, construct a quadratic B-spline curve between adjacent nodes to smooth the path, and solve the kinematic constraint problem during the robot's movement.
2. The robot path planning method based on the improved RRT-Connect algorithm according to claim 1, It is characterized in that In step 2, a dynamic restricted sampling area is set, and sampling is performed in the restricted sampling area. The dynamic restricted sampling area Range satisfies the following formula: Where N is the number of random tree expansions, K 1 , K 2 is the algorithm threshold.
3. The robot path planning method based on the improved RRT-Connect algorithm according to claim 1, It is characterized in that In step 4, the target biased adaptive step size strategy is adopted, and the adaptive step size S * The calculation formula is as follows: Where G(x) is the target bias function, and its value determines the target bias degree of node expansion. μ is the initial value of the function. is the function change rate coefficient, and the target bias function G(x) independent variable x satisfies the following formula: Where Δx is the increment of the independent variable, check() is the collision detection function, and the return value is 1 if the random tree collides with the obstacle, otherwise the return value is 0.
4. The robot path planning method based on the improved RRT-Connect algorithm according to claim 1, It is characterized in that In step 8, the greedy strategy is used to trim redundant nodes, and a quadratic B-spline curve is constructed to smooth the path. The specific process is as follows: First, the starting point q start and the target point q goal Directly connected, if there is no collision with obstacles, the line connecting the starting point and the target point is directly used as the path. If a collision occurs, the previous node of the end point is connected to the starting point q start Connected, follow the above steps, perform collision detection on the line connecting each node and the starting point in turn, until no collision occurs, and record the node as the new path node q new1 , and then the new path node q new1 As the new starting point, repeat the above process until a point that matches the target point q is found. goal The process ends when a new path node is directly connected without collision. start 、Target point q goal Connecting them in sequence will get the optimized path after clipping. The equation of the quadratic B-spline curve is as follows: where F i,2 (t) is the quadratic B-spline basis function, P i For the control point.
Citation Information
Cited By
Cable type underwater robot path planning system and method
CN119596997A
Multi-scene self-adaptive guide robot interaction method and system
CN120422235A
