Energy-Efficient Obstacle Avoidance via Iterative B-Spline Optimization for a Mobile Manipulator in Dynamic Environments

Abstract

This paper proposes an energy-efficient collision avoidance system for an autonomous mobile manipulator (AMM) in dynamic environments. The system enables safe obstacle avoidance and energy-efficient path generation. A B-Spline-based path optimization algorithm is developed, incorporating an energy cost function for planning collision-free, energy-efficient paths. The Bi-RRT algorithm generates an initial path, segmenting it at curvature maxima for energy-efficient optimization. The local path planning creates a path with seven control points, applying the same optimization for safe, low-energy avoidance of dynamic obstacles. A Model Predictive Control (MPC) system ensures precise path following. Experiments with a TM5M-900 robot validate the method's effectiveness in avoiding obstacles. Compared to Bi-RRT and Hybrid-RRT systems, it improves motion smoothness by 29.35% and reduces energy consumption by 16.84%, providing a more efficient solution for an AMM in dynamic environments.