Trajectory Tracking for Swing Phase of the Lower Limb Exoskeleton

In this paper, the dynamic model of swing phase of the lower limb exoskeleton is firstly established. Based on the proposed model, system identification experiments at each joint are conducted to determine unknown parameters and joint friction torques by tracking desired reference signals. The experimental data are processed by the least squares method. Another validation experiment is taken to verify the accuracy of the estimated model. Both linear and nonlinear controllers are designed. The linear controller is a proportional-derivative (PD) controller, and the nonlinear controller is a PD controller with gravity and friction torque compensation. Experiments show that the transient response of the tracking performance of the nonlinear controller is significantly better than that of the linear controller.