Efficient Path Planning of Robotic Arms Based on the Improved Informed-RRT* Algorithm

Authors:
Open Access
Publisher website
Google Scholar
Abstract
To address the limitations of the standard Informed-RRT* algorithm in manipulator path planning, including low initial path search efficiency, susceptibility to local optima in complex obstacle environments, and high path redundancy, this paper proposes an improved Informed-RRT* algorithm tailored for manipulator applications. First, we construct a phased adaptive sampling framework that separates the initial path search and path optimization stages. A target region constraint strategy is introduced, and the sampling confidence probability is dynamically adjusted based on the current search phase and real-time path quality. This design significantly enhances the efficiency of feasible path discovery while effectively preventing premature convergence to local optima. Second, an adaptive step size mechanism guided by gravitational–repulsive coordination is developed. This mechanism dynamically adjusts the extension step size according to the local obstacle distribution, reducing invalid sampling and increasing the number of effective sampling points while strictly ensuring obstacle avoidance safety, thereby accelerating both path search and optimization processes. Finally, a dichotomy-based dynamic boundary path smoothing strategy is integrated to generate smooth intermediate path points near obstacle boundaries. This strategy eliminates redundant inflection points and reduces path length while maintaining a safe distance from obstacles. The performance of the proposed algorithm is comprehensively verified through multiple sets of comparative experiments in both 2D grid maps and ROS-based manipulator simulation environments. The experimental results demonstrate that compared with the standard Informed-RRT* algorithm, the proposed method achieves a relative reduction of 77.17% in the average time to first initial path in complex environments. The path planning success rate increases from 21% to 95%, corresponding to an absolute increase of 74 percentage points and a relative increase of 352.38%. Additionally, the average path length is relatively reduced by 24.81%.
Funding Information
  • Electric Power Research Institute of State Grid Shanxi Electric Power Corporation (52053025000J)

This publication has 26 references indexed in Scilit: