A Robust Impedance Control Using Recurrent Fuzzy Neural Networks
Tsai-Jiun Ren · 2008
This paper presents a new adaptive impedance control based on a recurrent fuzzy neural networks (RFNN). The proposed control scheme includes two elements, a RFNN impedance nominal controller (RFNNINC) and a RFNN robust compensator (RFNNRC). The RFNNINC is developed to allow the linearized system performance to approximate the set impedance model accurately. The nonlinear term error between the system and linearized model uses the RFNNRC to compensate. Furthermore, when the system suffers external load and parameter variances, the RFNNRC can provide comparative force to resist the disturbances, allowing the entire system to be robust. Overall, the system is robust and has the desired impedance response. Some computer simulation results demonstrate the effectiveness of the proposed scheme for impedance control.