Robust Direct Adaptive Fuzzy Control of Flexible Joints Robots with Time-Varying Stiffness/Damping Parameters

Abstract:
In this article, we address the problem of controlling unknown flexible-joint robots with unknown time-varying stiffness and damping parameters. We propose a Robust Direct Adaptive Fuzzy Control (RDAFC) strategy that accommodates the dynamics anonymity and joints stiffness/damping variations. The RDAFC strategy relies on the synergy of the concepts of fuzzy logic approximation and the Sliding Mode Control (SMC). The fuzzy logic approximation relaxes the need for knowing the robot dynamics and the SMC accommodates the parameters variations. We also modify the RDAFC strategy to be suited to the KUKA Lightweight Robot (LWR) and propose a control strategy that can accommodates dynamics anonymity, uncertainty and joints elasticity variations. Experimental results are performed on a KUKA LWR moving in free space with its joints stiffness and damping vary with time in sine and cosine waveforms respectively. From the experiments, we can see that excellent tracking performance is obtained when using the RDAFC strategy despite the joints elasticity parameters time-variance and the robot dynamics unavailability.