Innovation Series: Advanced Science (ISSN 2938-9933, CNKI Indexed)

An Improved Coupling Strategy for Obstacle Avoidance in Dynamic Movement Primitives

Aosheng Sun

¹ University of Shanghai for Science and Technology, Shanghai, China

Corresponding Author: Aosheng Sun (sunaosheng23@163.com)

Abstract: Dynamic Movement Primitives (DMPs) are widely used in robotic trajectory generation and imitation learning due to their stability, parameter tunability, and generalization capability. However, most existing DMP-based obstacle avoidance methods rely on conventional artificial potential fields, which often suffer from trajectory oscillations, shape distortion, and loss of demonstrated features in complex environments. To address these issues, this paper proposes an obstacle-avoidance trajectory generation method for DMPs based on an improved artificial potential field. By incorporating an exponential distance attenuation function and a velocity-direction modulation mechanism into the coupling term, the proposed method achieves improved continuity and stability of the obstacle avoidance force in both spatial and directional domains, enabling adaptive local deformation of demonstrated trajectories. While preserving the original convergence property and modular structure of DMPs, the proposed approach significantly enhances trajectory smoothness and obstacle avoidance stability in both single- and multi-obstacle scenarios. Simulation results based on handwritten trajectory data demonstrate that the proposed method outperforms the conventional artificial potential field and steering-angle methods in terms of minimum error and root-mean-square error (RMSE), while better preserving demonstrated trajectory features.

Keywords: Dynamic movement primitives; Trajectory planning; Obstacle avoidance

References