An Efficient Method for Considering Time Correlated Noise in GPS/INS Integration

Jan Wendel, and Gert F. Trommer

Abstract:	In most integrated GPS/INS systems, a Kalman filter exploits the information provided by the GPS receiver in order to estimate the errors of the INS state and to calibrate the IMU. However, a Kalman filter does only provide optimal estimates, if process and measurement noise are white, which is often not the case. In this paper, the mechanism is clarified that causes the deterioration of performance or even divergence of a Kalman filter in the presence of time correlated noise: Time correlated measurement noise causes cross- correlations to build up between the measurement noise and the filter state vector, and time correlated process noise causes cross-correlations to build up between the process noise and filter state vector. As these cross- correlations are not considered in a 'conventional'. Kalman filter, the covariance matrix of the filter state vector is in error. From this insight, filter equations are derived which take into account this building up of cross- correlations, providing an efficient and flexible means of considering time correlated noise. However, for any technique of considering time- correlated noise, the knowledge of appropriate noise process models is mandatory. Often, e.g. for missile applications, the noise corrupting the inertial sensor data is caused mainly by the vehicle vibrations. Therefore, a method is developed for identifying models of this vibration induced noise during flight. Simulation results show the improvement that results if the vibrations are correctly taken into account by means of the derived filter equations, and prove that with the proposed identification method, the necessary noise process model parameters can be estimated despite the trajectory dynamics.
Published in:	Proceedings of the 2004 National Technical Meeting of The Institute of Navigation January 26 - 28, 2004 The Catamaran Resort Hotel San Diego, CA
Pages:	903 - 911
Cite this article:	Wendel, Jan, Trommer, Gert F., "An Efficient Method for Considering Time Correlated Noise in GPS/INS Integration," Proceedings of the 2004 National Technical Meeting of The Institute of Navigation, San Diego, CA, January 2004, pp. 903-911.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In