Classifications of Intentions for Controlling of a Multifunctional Knee-Ankle-Foot Orthosis

This paper describes motion intention classifiers which utilize reaction forces signals from heel and toe; and hip velocity information to predict subject’s intention. Those classifiers using Bayes method to predict (i) walk-to-stop, (ii) walking-speed-changing, and (iii) stop-to-motions. They are very accurate (most of them have accuracy rate higher than 90%) and a significant step in order to develop a multifunctional knee-ankle-foot orthosis.