State estimation of wheel chair motion by nonlinear Kalman filter

Abstract

This paper proposes a method for estimating the movement of a wheelchair using inertial sensors, even in the absence of GPS. A plant model was constructed by a kinematics model which includes wheelchair’s body section and two wheels, and constraint equations of differential type two-wheeled vehicle model. Simulating wheelchair tennis motion was performed and measured with an inertial sensor and optical motion capture. By applying the extended Kalman filter to the model described above, the wheelchair motion was accurately estimated for 20 seconds in duration.