Large -magnitude disturbance rejection and recovery based on human reflexes in an anthropomorphic bipedal locomotion system.

摘要

An anthropomorphic bipedal locomotion system is studied to investigate the system's robustness and susceptibility to loss of stability in the face of large magnitude disturbances such as slipping and tripping. The study is performed via mathematical simulation of a planar five-link biped model and closed-loop control system.;The equations of motion are formulated using the Euler-Lagrange differential equation and the Denavit-Hartenberg coordinate frame representation. The equations of impact are derived using the principles of linear and angular impulse and momentum. The equations are formulated and simulated to correctly model surface friction effects and slipping.;A nominal control system is formulated to accommodate all phases of locomotion including standing, starting, progression, and stopping. The biped is capable of progression over various surface grades with variable step length and progression velocity. Joint trajectory tracking is performed using the robust feedback linearization technique known as sliding mode control. Sub-optimal gait synthesis parameters are determined as the basis of an on-line gait optimization algorithm.;The basic capabilities of the nominally controlled bipedal locomotion system are demonstrated with simulation results. The susceptibility of the system to loss of stability after slipping and tripping is demonstrated.;A reflexive control algorithm is developed and integrated into the nominal control structure. The reflexive control successfully rejects the large magnitude disturbances of slipping and tripping. The performance of the reflexive control algorithm is demonstrated and the improvement in robustness of the locomotion system is quantified.