Abstract
The walking ability is very important in daily life. Lack of walking ability causes a decline in quality of life (QOL). The intelligently controllable walker (i-Walker) is a computer-aided walking support which recognizes environment and assists walking. In previous studies we has conducted several experiments and identified several problems. And we found that if there is mismatch between intent of user's operation and the control of i-Walker, it prevents smooth walking of users. To solve this problem, we developed a sensor walker (Pre-i-Walker4) and a prediction method by using support vector machine. Gait experiments were conducted by healthy users to evaluate the developed prediction model to predict stop, walk, turning left and right, and emergency.
