01337394

Component

This paper aims to develop an enhanced tightly-coupled LVT system by integrating data from GPS, low-cost dead-reckoning (DR) sensors and a Digital Elevation Model (DEM). Three different configurations are considered: (1) integrating GPS with low-cost DR sensors resulting in an integrated GPS/DR system, (2) integrating GPS with a Digital Elevation Model (DEM) resulting in an integrated GPS/DEM system and (3) integrating GPS with both DR sensors and a DEM resulting in an integrated GPS/DR/DEM system. An Extended Kalman Filter (EKF) is employed to integrate data in each of the three integrated systems. The accuracy performance of each of the three LVT systems has been evaluated using both real-world and simulated data. Simulated data have been created to replicate a relatively harsh operational environment where the performance of a LVT system is expected to degrade rapidly with low GPS satellite visibility. The results suggest that the GPS/DR/DEM system, as expected, outperforms the other two systems (GPS/DR and GPS/DEM). This is because the positioning error in the GPS/DR/DEM system is compensated by the GPS measurements, DR measurements and DEM height data even with only one satellite in view. If no satellite is available (e.g., within a tunnel), the positioning error in GPS/DR/DEM is reduced by DR measurements and DEM height data. The accuracy offered by the GPS/DR/DEM system has been found to be 1.83m (95%) for the GPS dataset collected in Nottingham (UK) and the maximum horizontal error at the end of 30sec continuous GPS outage has been found to be 6.6m (with three satellites in view) and 31.3m (with only one satellite in view). The overall cost of the GPS/DR/DEM system would be low given that the only additional sensor needed here is a low-cost gyroscope along with a GPS receiver. Therefore, the LVT system has capability to satisfy the accuracy requirements of many intelligent transport systems operating in dense urban areas and has high potential to be implemented by the industry.

01329018

11-2731

English

22p

2011

Transportation Research Board 90th Annual Meeting

DVD

Figures (3) ; References (21) ; Tables (5)

Accuracy; Automatic vehicle detection and identification systems; Costs; Data quality; Dead reckoning; Global Positioning System; Ground vehicles; Kalman filtering; Mathematical models; Position fixing

Elevation models

Data and Information Technology; Highways; Vehicles and Equipment; I91: Vehicle Design and Safety

Transportation Research Board Annual Meeting 2011 Paper #11-2731

TRIS, TRB

Feb 17 2011 6:16PM