Vehicle positioning in urban environments using particle filtering-based global positioning system, odometry, and map data fusion

Abstract

This article presents a new method for land vehicle navigation using global positioning system (GPS), dead reckoning sensor (DR), and digital road map information, particularly in urban environments where GPS failures can occur. The odometer sensors and map measure can be used to provide continuous navigation and correct the vehicle location in the presence of GPS masking. To solve this estimation problem for vehicle navigation, we propose to use particle filtering for GPS/odometer/map integration. The particle filter is a method based on the Bayesian estimation technique and the Monte Carlo method, which deals with non-linear models and is not limited to Gaussian statistics. When the GPS sensor cannot provide a location due to the number of satellites in view, the filter fuses the limited GPS pseudo-range data to enhance the vehicle positioning. The developed filter is then tested in a transportation network scenario in the presence of GPS failures, which shows the advantages of the proposed approach for vehicle location compared to the extended Kalman filter.