High accuracy tightly-coupled integrity monitoring algorithm for map-matching

Abstract

A map-matching algorithm employs data from Global Positioning System (GPS), a Geographic Information System (GIS)-based road map and other sensors to first identify the correct link on which a vehicle travels and then to determine the physical location of the vehicle on the link. Due to uncertainties associated with the raw measurements from GPS/other sensors, the road map and the related methods, it is essential to monitor the integrity of map-matching results, especially for safety and mission-critical intelligent transport systems such as positioning and navigation of autonomous and semi-autonomous vehicles. Current integrity methods for map-matching are inadequate and unreliable as they fail to satisfy the integrity requirement due mainly to incorrect treatment of all the related uncertainties simultaneously. The aim of this paper is therefore to develop a new tightly-coupled integrity monitoring method for map-matching by properly treating the uncertainties from all sources concurrently. In this method, the raw measurements from GPS, low-cost Dead-Reckoning (DR) sensors and Digital Elevation Model (DEM) are first integrated using an extended Kalman Filter to continuously obtain better position fixes. A weight-based topological map-matching process is then developed to map-match position fixes onto the road map. The accuracy of the map-matching process is enhanced by employing a range of network features such as grade separation, traffic flow directions and the geometry of road link. The Receiver Autonomous Integrity Monitoring (RAIM) technique, which has been successfully applied to monitor the integrity of aircraft navigation, is modified and enhanced so as to apply it to monitor the quality of map-matching. In the enhanced RAIM method, two modifications are made: (1) a variable false alarm rate (as opposed to a constant false alarm rate) is considered to improve the fault detection performance in selecting the links, especially near junctions. (2) a sigma inflation for a non-Gaussian distribution of measurement noises is applied for the purpose of satisfying the integrity risk requirement.The implementation and validation of the enhanced RAIM method is accomplished by utilising the required navigation performance parameters (in terms of accuracy, integrity and availability) of safety and mission-critical intelligent transport systems. The required data were collected from Nottingham and central London. In terms of map-matching, the results suggest that the developed map-matching method is capable of identifying at least 97.7% of the links correctly in the case of frequent GPS outages. In terms of integrity, the enhanced RAIM method provides better the fault detection performance relative to the traditional RAIM.