阈值引导树优化的RRT*机械臂路径规划算法研究OA

In allusion to the problems of strong randomness,poor guidance,and low sampling efficiency in the traditional rapidly-exploring random tree(RRT)algorithm,an improved RRT* algorithm based on threshold-guided tree is proposed.In this algorithm,node values are evaluated to filter out high-value nodes,and the directional growth of the random tree is implemented by combining with dynamic threshold iteration,so as to reduce the search randomness.In order to prevent getting stuck in local exploration deadlocks,a lightweight restart mechanism is introduced.An adaptive sampling strategy is used to efficiently allocate sampling resources and enhance the goal-oriented nature of sampling.A direct connection strategy is introduced to accelerate the algorithm convergence.The cubic B-spline smoothing on the generated path is conducted.The improved RRT* algorithm is compared with related algorithms.The simulation experimental results show that the improved RRT* algorithm is improved in terms of path length,number of path nodes,number of iterations and node utilization,verifying the feasibility and effectiveness of the improved algorithm.