Accounting for sensor drift in miniature, wireless inertial measurement and positioning systems: an extended Kalman Filtering approach

Abstract

This work describes the design and development of a location aware, wireless inertial measurement system incorporating Kalman Filtering based real-time correction for sensor drift. The Tyndall prototyping mote is employed as a wireless sensor networking platform and its suitability as a flexible architecture for a wide range of applications is highlighted. The system comprises of an Inertial Measurement Unit (IMU) with 3 axis gyroscopes, accelerometers and magnotometers and a newly developed 25mm Global Positioning Satellite (GPS) transceiver interface layer. A methodology for fusioning GPS and IMU data is presented highlighting how an extended Kalman filter can be employed to re-calibrate the IMU in realtime, correcting for sensor drift. A Labview based visualisation tool is described enabling real-time analysis of GPS, IMU, Kalman Filter and correction data. Finally some initial experimental results are presented highlighting the effectiveness of the system.