DMPs-based skill learning for redundant dual-arm robotic synchronized cooperative manipulation

Abstract Dual-arm robot manipulation is applicable to many domains, such as industrial, medical, and home service scenes. Learning from demonstrations a highly effective paradigm for robotic learning, where learns human actions directly can be used autonomously new tasks, avoiding the complicated analytical calculation motion programming. However, learned skills are not easy generalize cases special constraints varying relative distance limitation of end effectors human-like cooperative manipulations exist. In this paper, we propose dynamic movement primitives (DMPs) based learning framework redundant dual-arm robots. The method, with coupling acceleration term DMPs function, inspired by transient performance control Barrier Lyapunov Functions. additional calculated on constant joint limitations object-approaching actions. addition, integrate generated in space solution complete manipulation. Simulations undertaken Matlab Gazebo environments certify effectiveness proposed method.