Adaptive motion/force control of robots with input nonlinearities via fuzzy logic system

In this paper, an adaptive motion/force control scheme is proposed via fuzzy logic system for solving the problem of input backlash in the practical application of humanoid robot manipulation. Compared with the existing literatures ignoring input nonlinearities, an smooth backlash inverse is firstly employed to deal with the lines-segments effect from backlash nonlinearity. Secondly, a decentralized control configuration is proposed to make the motion and the internal forces converge to the predetermined values at the same time. Moreover, the stabilities of all signals in the robotic system are guaranteed by the Lyapunov second method. Finally, simulation and experiment illustrate the effectiveness of the proposed control scheme.