Application of Real-time Kalman Filter with magnetic calibration for MEMS sensor in attitude estimation

This paper addresses the development of a high precision low-cost Inertia Measurement Unit (IMU) where MEMS-based accelerators, gyroscopes, magnetic sensors and an extended Kalman filter are used in a real-time signal processing system. Integral operation of inertial sensor signals leads to velocity and position estimation drift that increases with time. The drift has been eliminated by our algorithm of Kalman Filtering technique with geomagnetic information. The algorithm are tested in Matlab environment and implemented on a PIC microcontroller to estimate the attitude of an object in space. MATLAB/Simulink Embedded tools and compiler C30/MPLAB are used to design, compile and download directly into the target. The processing rate of 100Hz with real-time floating point calculation can be achieved for the extended Kalman algorithm using a Microchip 16 bit dsPIC33f256.