A Kalman filter for integrating dead reckoning, map matching and GPS positioning

A Kalman filter has been developed to integrate the three positioning systems (differential odometer dead reckoning, map matching, and Global Positioning System or GPS) used in the Automatic Vehicle Location System (AVL 2000) being designed and developed in the Department of Surveying Engineering at the University of Calgary. The system is being targeted for on road applications and incorporates a digital map. The filter has been designed to take into account uncertainties via covariance matrices. In wide-open spaces GPS positioning will dominate, while in zones where the GPS signal is obstructed, dead reckoning will be used as interpolation between GPS position fixes. Simulation studies and covariance analyses have been performed on a test route located in a sector of the city of Calgary.>