Bio-inspired control of lower limb exoskeleton using a central pattern generator

This paper, we present a new framework for lower extremity exoskeleton to generate on-line trajectory and get in sync with the environment based on a simplified neural-skeletal-muscle system. In contrast to non-autonomous movement representations like splines, because of the introduction of the concept of central pattern generator (CPG) in biology, this system is autonomous dynamical system, i.e. a system of differential equations with at least one limit cycle attractor. This neural-skeletal-muscle system consists of 7 segments and 18 virtual muscles, and controlled by a CPG network composed of 6 pair of neural oscillators, and mechanisms for processing and transporting sensory and motor signals. We evaluate the system with a humanoid robot simulation and an actual lower limb exoskeleton. Through the test, we found that locomotion emerged as a stable limit cycle that was generated by the global entrainment between the skeletal-muscle system, the neural system, and the environment.