Data fusion techniques and positioning estimation for land vehicle navigation systems: an overview

Abstract

This article describes the problem of real-time estimate of the position of a vehicle for use in terrestrial navigation systems. After describing the context of the application and give a definition of the problem, we review the mathematical framework and the technologies involved in designing positioning systems. We compare the performance of more popular data fusion approaches and give a view of their limitations and capabilities. Then check for the robustness of the positioning systems when one or more sensors are defective. It describes how the positioning system can be made more robust and adaptable, taking into account the occurrence of failure of sensors. Finally, it goes a step further and explores the possible architectures for collaborative positioning systems, where multiple vehicles interact by exchanging data to improve the estimate of its own location. The article finishes with some conclusions about future developments in this field.