Electric Wheelchair Controlled by Human Body Motion -Classification of Body Motion and Improvement of Control Method-

Abstract

This paper classifies human body movements when an electric wheelchair was controlled using a Human Body Motion Interface (HBMI) by a Self-Organizing Map (SOM) and proposes control based on classification results. The Human Body Motion Interface (HBMI) uses body movement following voluntary motion. This study focuses on electric wheelchair control as an application of the HBMI. The viability of the HBMI was confirmed using Center Of Weight (C.O.W.) from pressure distribution information on backrest in the wheelchair to control it. If body movement concentrated on a single point at C.O.W. in pressure distribution, a problem occurred because the system would recognize even different body-movement patterns as the same movement. We call body movement taking the same C.O.W. even if it has a different body-movement pattern movement confusion. We solve the movement confusion problem and enhance wheelchair control, classifying body movement using the SOM and reflecting this classification result to improve wheelchair control. Experimental results showed that movement confusion is solved and wheelchair control improved.