Learning Controllers for Lower-limb Prosthetic Devices


Professor Timothy Bretl of the Aerospace Engineering and Navid Aghasadeghi of ECE both from the University of Illinois have developed a method for learning controller parameters for lower-limb prostheses using only observed unimpaired trajectories and the physical characteristics of amputees.

This technique is an improvement over the usual trial and error method that is currently used in the fitting of lower-limb prosthetic devices.

Patient characteristics are applied in an algorithm to automatically learn controller parameters for level-walking at a desired cadence, stair/ramp ascent and other modes of locomotion.