Pedestrian motion based inertial sensor fusion by a modified complementary separate-bias Kalman filter

This paper presents a modified complementary separate-bias Kalman filter for orientation determination of pedestrian motions by using a inertial measurement unit (IMU) module, which contains gyroscopes, accelerometers and magnetometers as an Attitude and Heading Reference System (AHRS). The filter consists of two main functions: the complementary separate-bias Kalman filtering avoids the modelling of pedestrian motions and fuses the sensed data; the magnetic disturbance detection and minimization provides robustness and stability when the sensor module is experiencing local magnetic disturbances. Test case includes stairs climbing indoors and long-distance walking outdoors. In both case the filter is able to provide stable orientation data and minimize the impact of local magnetic field disturbance.