GUIDANCE ESTIMATE, BASED ON KALMAN FILTER, USING UNIT INERTIAL MEASUREMENT WITHOUT MAGNETOMETER

Abstract

This work is based on an already developed electronic system, comprised by inertial sensors, without magnetometer. This is unable to provide an orientation measure that allows obtaining gait parameters, since the signals to gather from it are: angular velocity and linear acceleration. To add this functionality to the system, several algorithms developed in the literature were analyzed and implemented. A comparison between the previous implementations is made from experimental signals. These signals are registered from rotations with different values of angle with respect to the 3 coordinate axes. Among the algorithms consulted are the straight integration of the gyroscope signal and the fusion of sensors. In this context, some problems were found related to the way to obtain the angles from the output signals of the accelerometer and the gyroscope. Therefore, in this work it is proposed an implementation, based on the Kalman filter, to obtain the orientation without using a magnetometer and that minimizing the deviation produced by the integration of the gyroscope signal. This algorithm was applied to dynamic equilibrium tests that are carried out in the health institution.