Home // International Journal On Advances in Networks and Services, volume 13, numbers 3 and 4, 2020 // View article

Evaluation of Vehicle Position Estimation Method Combining Roadside Vehicle Detector and In-vehicle Sensors

Authors:
Shunya Yamada
Yousuke Watanabe
Hiroaki Takada

Keywords: Sensor fusion; Position estimation; Communication delays; Intelligent transportation system.

Abstract:
Properly managing merging at junction is important in improving highway traffic safety and traffic flow. Accurate vehicle positions and velocities are necessary in realizing proper management, but existing sensors have both advantages and disadvantages. Roadside vehicle detectors are very accurate, but only available at fixed points. By contrast, in-vehicle Global Navigation Satellite System (GNSS) sensors can be used anywhere except in tunnels, but are less accurate. However, these sensors can compensate for each other's week points. In this paper, we propose a vehicle position estimation method that combines roadside vehicle detector and in-vehicle sensors. This method is used to gather data from the roadside vehicle detector and in-vehicle sensors via different wireless networks. It then applies Kalman filters to calculate the accurate position and velocity. When exchanging information over wireless networks, communication delays occur and the data arrival sequence is not guaranteed. Our method can retroactively calculate the vehicle position in the presence of delays below a maximum acceptable threshold. This study evaluates in several simulations using the data created from the GNSS position error and communication delay models. We also perform the evaluation using a dataset created based on the actual vehicle information and communication delays in the simulations. The simulation results show that our method can more accurately estimate the vehicle positions compared to that using data from either sensor alone. Moreover, our method is more suitable for managing traffic and controlling merging at junctions.

Pages: 82 to 93

Copyright: Copyright (c) to authors, 2020. Used with permission.

Publication date: December 30, 2020

Published in: journal

ISSN: 1942-2644