Heuristic RRT mechanical arm motion planning method based on target deviation optimization

Abstract

The invention belongs to the technical field of robot path planning and generation in a three-dimensional space, and particularly relates to a heuristic RRT mechanical arm motion planning method basedon target deviation optimization. The method comprises the following steps: determining a starting point q<start>, and storing the q<start> in q<nodes>; randomly sampling a point q<rand> in a space in a full map, and searching a nearest point in the q<rand> where q<nodes> reach as q<near>; controlling a random point generation direction and distance to allow the random point to move forward to anode q<new> in the direction from q<near> to q<rand> at a step length delta by utilizing a target deviation factor; carrying out collision detection in the forward process, and returning to A2 if collision is detected; if no collision is detected, storing the q<new> in the q<nodes>, and turning to A4; repeating A2 to A3 according to the updated q<nodes> until the obtained latest node q<global> meets the relational expression shown in the description; regarding the latest node q<global> as a target point, and storing the q<new> into the q<nodes>; and reversely searching at q<nodes> according tothe father-son relationship of each node to find a planned path. According to the invention, the length of the searching path can be reduced, and the phenomenon of falling into a local optimal valueor shaking near an obstacle is avoided.

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