Control methods for powered assistive devices for individuals with mobility impairments: compensating for disrupted physiological control loops

Abstract

Physiologic control of purposeful movement requires continuous interactions between the sensory and motor systems. When physiological control loops are disrupted, as is the case in individuals with limb amputations or spinal cord injuries, purposeful movement can be difficult to achieve. With recent advances in robotic technology, assistive devices replacing some of the lost functions have begun to emerge in both research and commercial settings. This dissertation describes control approaches for such devices, with the goals of compensating for impaired physiological functions in individuals with mobility impairments while maximizing the use of retained physiological functions. Specifically, this dissertation describes the following control methods: (1) a machine learning algorithm using surface electromyography (EMG) for volitional control of powered prosthetic devices for individuals with limb amputations, (2) a high-level step trigger controller using an inverted pendulum model for lower limb exoskeletons for gait restoration in individuals with spinal cord injuries (SCIs), and (3) a cooperative controller that combines the lower limb exoskeleton with the user’s own muscles using functional electrical stimulation (FES). Each control method was tested on multiple subjects for whom the controller was developed. The volitional control method using EMG was tested on three subjects with transfemoral amputations. The step trigger controller for lower limb exoskeletons was tested on five subjects with motor-complete spinal cord injuries. The cooperative controller combining FES with powered lower limb exoskeletons was tested on three subjects with motor-complete spinal cord injuries who were FES-responsive. In all cases, results indicate that the proposed control methods are able to provide functional benefits to individuals with mobility impairments.