Scholarlink

<p>An improved RRT algorithm incorporating dynamic cone constraints and adaptive step size adjustment was proposed to address the limitations of conventional Rapidly-exploring Random Tree (RRT) algorithms in robotic arm path planning, including excessive randomness in sampling, slow convergence, and non-smooth trajectories. The method constructed a dynamic conical model to confine sampling within an adaptively adjusted conical region. A step size regulation mechanism based on obstacle distance feedback dynamically optimized exploration increments. Greedy strategies combined with cubic B-spline curve optimization were applied to ensure global path optimality and local smoothness. Comparative MATLAB simulation results demonstrated that compared with the standard RRT, the improved RRT is 84.76% lower in average planning time (2D) and 73.03% lower (3D), and it also is 20.13% lower in the average path length (2D) and 25.20% lower (3D), as well as having 46.77% fewer path nodes (2D) and 66.94% fewer (3D). In 6-DOF robotic arm scenarios, it achieves 64.04% faster planning, 34.90% shorter paths, and 72.38% fewer nodes. Joint angular velocity and acceleration exhibit smooth, continuous dynamics throughout motion trajectories with no discontinuities. This research effectively improves the inefficiencies of RRT in complex scenarios and significantly optimizes real-time high-dimensional motion planning. Physical validation via a comprehensive experimental setup in the Robot Operating System (ROS) confirms the algorithm’s advantages over conventional RRT.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An improved RRT-based path planning approach with dynamic cone angle guidance for robotic manipulator obstacle avoidance

  • 期刊论文

  • 发表日期：

    2026年1月17日

查看全文
  • Xifeng Liang,
  • Feng Yu,
  • Wenshuo Gao,
  • Baoguo Yao

摘要

An improved RRT algorithm incorporating dynamic cone constraints and adaptive step size adjustment was proposed to address the limitations of conventional Rapidly-exploring Random Tree (RRT) algorithms in robotic arm path planning, including excessive randomness in sampling, slow convergence, and non-smooth trajectories. The method constructed a dynamic conical model to confine sampling within an adaptively adjusted conical region. A step size regulation mechanism based on obstacle distance feedback dynamically optimized exploration increments. Greedy strategies combined with cubic B-spline curve optimization were applied to ensure global path optimality and local smoothness. Comparative MATLAB simulation results demonstrated that compared with the standard RRT, the improved RRT is 84.76% lower in average planning time (2D) and 73.03% lower (3D), and it also is 20.13% lower in the average path length (2D) and 25.20% lower (3D), as well as having 46.77% fewer path nodes (2D) and 66.94% fewer (3D). In 6-DOF robotic arm scenarios, it achieves 64.04% faster planning, 34.90% shorter paths, and 72.38% fewer nodes. Joint angular velocity and acceleration exhibit smooth, continuous dynamics throughout motion trajectories with no discontinuities. This research effectively improves the inefficiencies of RRT in complex scenarios and significantly optimizes real-time high-dimensional motion planning. Physical validation via a comprehensive experimental setup in the Robot Operating System (ROS) confirms the algorithm’s advantages over conventional RRT.