KALMAN ADAPTIVE FILTER RESEARCH FOR ACCELEROMETER ON THE ELECTRIC WHEELCHAIR

Abstract

This paper presents a method of processing tilt angular accelerated signal of an accelerometer mounted on an electric wheelchair to alert the fall phenomenon during movement. The accelerometer is an ADXL345 sensor which generates accelerated signal in three X, Y, Z axis. This signal is passed through the Kalman filter to accurately estimate the angular acceleration regardless of the impact of system noise and measurement noise. In addition, the crash warning system during moving process is introduced in this paper. The simulation results implemented by Matlab software and experiment results show that the Kalman filter contributes the accurate measurement signal with low oscillation that provides reliability for the electric wheelchair control system.