Adaptive locomotion of lower limb exoskeleton based on oscillators and frequency adaptation

摘要

In this paper, we focus on the topic of adaptive control of the trajectory of the lower limb exoskeleton, and our ultimate goal is to establish a design principle of a controller in order to achieve natural human-like locomotion. We suggest dynamical movement primitives as a central pattern generator (CPG) of lower limb exoskeleton, this approach combines nonlinear oscillators, i.e. dynamical system that exhibit limit cycle behavior, this system can spontaneously generates trajectories for the robot. And by introducing adaptive factor, our control system is mathematically capable of learning the high level features (frequency, envelope, etc.) of a periodic input signal. Besides, we introduced foot pressure feedback into the system, and designed related hardware. Numerical simulations and experimental implementation on a physical robot demonstrate the effectiveness of the proposed locomotion controller.