Positioning and Tracking in Multipath Environment for Vehicular Scenarios

No Thumbnail Available

URL

Journal Title

Journal ISSN

Volume Title

Sähkötekniikan korkeakoulu | Master's thesis

Date

2021-12-13

Department

Major/Subject

Autonomous Systems

Mcode

ELEC3055

Degree programme

Master's Programme in ICT Innovation

Language

en

Pages

60+8

Series

Abstract

Vehicle-to-everything (V2X) is the critical enabling technology for the construction of the future intelligent transportation system. Providing accurate and real-time location information of the connected vehicles is the prerequisite in the V2X implementation. The existing radio-based positioning solutions either have insufficient positioning accuracy or are subject to the signal intensity problem. Due to several properties: high carrier frequencies, large bandwidths, a large number of antenna arrays, direct device-to-device communication, and network densification, 5G is a promising candidate to provide decimeter level positioning at an economical deployment cost. In this project, we propose a versatile and high-accuracy vehicular position estimator using 5G wireless radio technology with a single base station regardless of the presence of line-of-sight signal transmission. Two extended Kalman filters (EKF) were utilized in the system for different tracking purposes. Specifically, the first EKF was formulated to computation efficiently track the channel parameters, while the second EKF was used for vehicle position tracking. We implement a triangulation positioning module to calculate the vehicle position and orientation from the angles and path lengths information. In the meanwhile, the triangulation positioning module connects the two separate EKFs. Besides, an abrupt change detection method was developed to address the significant break problem of the propagation channel with the movement of the vehicle in the network. Furthermore, the proposed 5G vehicle position estimator was extended with IMU integration to meet higher positioning requirements. Finally, we validate and evaluate the performance of our positioning system in several simulation experiments using MAE, RMSE, and cumulative density function metrics. The numerical result demonstrates that decimeter positioning accuracy can be achieved using the designed 5G positioning system even when the line-of-sight propagation path is blocked. Higher vehicle positioning accuracy is possible when integrating the IMU sensor.

Description

Supervisor

Jäntti, Riku

Thesis advisor

Vinogradova, Julia

Keywords

vehicle positioning, extended kalman filter, 5G, antenna, triangulation, vehicle-to-everything

Other note

Citation