Heuristic rrt manipulator motion planning method based on goal-biased optimization

Abstract

The present invention belongs to the robot path planning in the three -dimensional space and generating technology. It involves a method of inspiration RRT mechanical arm motion based on the target bias.Including the following steps: Determine the starting point Q start , To Q Q start Store to Q nodes Middle; in the space, the full map random sampling point Q rand , Find Q nodes Arrival Q rand One of the most recent points as Q near ; near To Q rand The direction of the step -long Δ is advanced, and the target bias factor controls the random point to generate direction and advances to node Q new , Perform a collision detection in the process. If the collision is detected, it returns A2; if the collision is not detected, Q new Store to Q nodes Middle and turn to A4; according to the updated Q nodes , Repeat A2 ~ A3 until the latest node Q goal Satisfaction | Q new Molten Q goal | <Error, deemed to reach the target point, and Q new Start in Q to Q nodes Middle; in Q nodes According to the father -son relationship of each node, the reverse search finds the planning path.The present invention can reduce the length of the search path and avoid the phenomenon of being in a local optimal value or shaking near obstacles.

Description

technical field [0001] The invention belongs to the technical field of robot path planning and generation in three-dimensional space, and in particular relates to a heuristic RRT manipulator motion planning method based on target bias optimization. Background technique [0002] Motion planning can be divided into path planning and trajectory planning. Path planning focuses on the generated path, and trajectory planning is to give the path time information. For the manipulator, its end effector is a technical object. In path planning, the main technology is the path point tracking problem. In trajectory planning, it is based on The job task requires to realize the practical application of kinematics inverse solution (position and orientation inverse solution, velocity and acceleration inverse solution), so a systematic motion planning is of great significance for solving practical problems. [0003] The RRT algorithm is a fast path planning algorithm based on random sampling...

AI Technical Summary

Problems solved by technology

Among them, LaValle introduced probability-based sampling, which improves the search efficiency of the basic RRT algorithm, but at the same time it is easy to fall into the local minimum

Karaman proposed the RRT algorithm with the optimal path, the planning path has high stability and is close to optimal, but it takes a lot of time

The above techniques cannot completely avoid the phenomenon of local optimum or oscillation near obstacles, and the reliability of the algorithm is not strong for the manipulator

At the same time, motion planning is separated from the technology, most of the algorithms stay in two-dimensional space, and the feasibility of practical application in three-dimensional space cannot be guaranteed

Images