Cascading filtered DTS data into a loosely-coupled GPS/INS system

This paper addresses the integration of a subordinate Kalman filter processing terrain data into a Kalman filter integrating Global Positioning System (GPS) data and Inertial Navigation System (INS) data with limited computing resources. The study shows that modeling the difference between the horizontal position estimates supplied by the Digital Terrain System (DTS) and the integrating Kalman filter is sufficient for accurate solution integration. The two new error states required to model this difference are modeled as random walks driven by variable process noise. Augmenting the GPS/INS integrating Kalman Filter with the horizontal position offset states requires minute changes to the existing filter mechanization. Performance measurements of position, velocity and platform tilt error using Monte Carlo simulations show stable and acceptable navigation accuracy. An alternative integration of subordinate solutions scheme is also suggested and presented