Online adaptation of optimal control of externally controlled walking of a hemiplegic individual

Bipedal walking is a quasi-cyclic activity where the movements of the ipsilateral leg and contralateral leg are time shifted for 50 percent of the gait cycle. Hemiplegia, a most common consequence of cerebro-vascular accident, results with unilateral disability; thereby the pattern of the paralyzed leg is modified, and in many cases prevents or greatly decreases the ability to walk normally. One possible method of restoring the movements of the paralyzed leg is functional electrical stimulation (FES), that is, patterned electrical activation of sensory-motor systems of the paralyzed leg. We developed a new method that combines several techniques that have been tested for control of FES: 1) optimal control based on a customized model of the individual using the system, 2) modeling of the plant as a nonlinear oscillator, and 3) triggering of the oscillator from the sensor that is representative of a specific phase during the gait cycle of the healthy leg. The optimal control is used to create the basic pattern for stimulation; the oscillator, modulated by the sensory measurements, brings to the control information related to the rhythm.