This project will develop a novel control system to automatically regulate posture and actively restore balance to users of neuroprostheses for standing after spinal cord injury. Current standing neuroprostheses utilizing functional neuromuscular stimulation (FNS) only provide support and prevent collapse by stiffening the lower extremities through continuous supramaximal stimulation of the knee, hip and trunk extensors. They include no mechanism to actively maintain balance or compensate for disturbances, and rely on the upper extremities to make the postural corrections required to remain balanced in the upright position. This project will address the shortcomings of currently available FNS standing systems by developing a sensor-driven """"""""artificial vestibular system"""""""" that will actively monitor posture, anticipate perturbations to balance and automatically modulate stimulation to keep the user upright. This will be accomplished by combining innovative feed-forward, feedback and adaptive control techniques at multiple joints and in three dimensions. A small number of simple, but information-rich, body-mounted sensors will capture the actions of the torso as well as the lower extremities. Dynamic stability will be achieved by using accelerations of the trunk to predict and rapidly respond to disturbances in a feed-forward manner. Static stability will be achieved by regulating center of pressure within the base of support using feedback control, and adaptive algorithms will be applied to compensate for fatigue. A model-based approach to controller development will be adopted that relies on computer simulation, optimization and performance verification prior human testing. This new control system should reduce reliance on the upper extremities while standing with FNS, thus advancing the goal of providing neuroprosthesis users with freer use of their hands to manipulate objects in the environment by automatically maintaining balance in the presence of intrinsic and extrinsic disturbances.
Showing the most recent 10 out of 22 publications
