Identification of Exoskeleton Dynamics Parameters and Control Based on Series Elastic Actuator

Abstract

The lower limb exoskeleton is used for rehabilitation, assistance and maturing. Taking an exoskeleton based on a serial elastic actuator as an object, a method of dynamic parameter identification is proposed. Firstly, the modular SEA structure is introduced, and the exoskeleton is modeled by using the Modified Denavit-Hartenberg (MDH) method. Then, the exoskeleton dynamic model is derived and linearized. Secondly, the excitation trajectory required for exoskeleton parameter identification is designed, which is optimized by using the genetic algorithm. Then, the exoskeleton is tested by using the excitation trajectory to decouple the dynamic parameters. Finally, the exoskeleton impedance control algorithm based on dynamic compensation is designed and compared with the algorithm without dynamic compensation. The results show that the parameters obtained from the identification algorithm and the control algorithm are effective.