Decentralized integral sliding mode control for modular and reconfigurable robot with harmonic drive transmission based on joint torque estimation

This paper presents a decentralized integral sliding mode control method for a modular and reconfigurable robot (MRR) with harmonic drive (HD) transmission. Unlike known methods that rely on joint torque sensing, this paper addresses the problems with a joint torque estimation technique using only position measurements. First, a high-fidelity HD model is presented to estimate the joint torque which is used to formulate the dynamic model of the MRR. Second, based on only local dynamics information of each module, a decentralized integral sliding mode control scheme is proposed to reduce the chattering effect of the controller and compensating the model uncertainty. The asymptotic stability of the system is proved using the Lyapunov theory. Finally, simulations are performed for a 3-DOF MRR to study the effectiveness of the proposed method.