Trajectory-Tracking of Nonlinear Biped Robot System Based on Adaptive Fuzzy Sliding Mode Control

Fuzzy theory used to control biped robot is a new research direction recently. Transfer function and state space are the way of traditional control in biped robot. But it can't handle nonlinear, uncertainty, or time variant system. Fuzzy theory use "IF∼THEN"-rules to state the relationship about input and output of control system. It does not need to modulate linear controller's parameters such as Kp , Ki, Kd, and it can handle nonlinear system. It is important for biped robot because linear system makes the robot stiff, but fuzzy can implement well. In this paper, we plan the trajectories of six joint angles and use fuzzy sliding model controller to track the trajectories. The difference between trajectory tracking and Lagrange method is the simplicity for computing by trajectory tracking. The sliding mode can control the biped to enter the stable state quickly due to sliding mode utilizing two controllers. We implement our experiment using MSC visualNastran 4D and MATLAB/simulink . The MSC-visualNastran-4D software can include the real physics parameters of the sticks and the motors. The simulation movement resulting of biped walking will be much more closer to real biped robot rather than using MATLAB only .In addition, the information or input parameters that the robot needs are more useful. ©2007 IEEE.