Skip to Main Content
This paper focuses on the design and test results of an adaptive Kalman filter (KF) estimator for fusing data from a pair of Real-Time Kinematic (RTK) Global Positioning Systems (GPS)s and an Inertial Measurement Unit (IMU) in order to estimate not only the position, velocity, and attitude of a vehicle in 3-dimension but also the IMU calibration parameters. Since GPS systems sometimes lose their signal and receive inaccurate position data, the self-tuning filter estimates also the covariance matrix associated with the GPS measurement noise, so that the KF filter is continually “tuned” as well as possible. The observability analysis of the linearized system is discussed in the paper and the result shows that the system is locally observable if the line connecting the two GPS antennas is not collinear with the vector of total acceleration, i.e., the sum of gravitational and inertial accelerations. Finally, test results obtained from a mobile robot moving across uneven terrain are presented.