Broad fuzzy neural control using impedance learning

This work proposes a novel control strategy based on broad fuzzy neural network (BFNN) by using impedance learning, which is subjected to contact with the unknown dynamic environment. Compared with the original fuzzy neural network, this framework is provided the prominent feature by taking the advantage of broad learning system (BLS) to approximate the unknown dynamic model. Aiming at offering a compliance contact scheme, this paper introduce the impedance learning to establish the robot-environment interaction model. Also, a stable controller, which is able to tackle the problems related to the state constrain, is designed through Barrier Lyapunov Function (BLF). The proposed method can achieve the favourable tracking action while guaranteeing the stability of closed-loop system. In the end, simulation study is performed to verify the effectiveness of BFNN with a two-DOF manipulator.