In order to design a fast and accurate short-distance localization system to supplement the gap of Global Position System (GPS) in a toll station, in this paper, an IEEE 802.11p module driver is modified to provide channel state information (CSI), and the signal processing techniques are leveraged in both time-domain and frequency-domain to mitigate multi-path effect. An effective CSI vector provided by a roadside unit (RSU) transceiver (RX) array is treated as the fingerprint of a transmitting spot, the spatial diversities of the multiple antennas are leveraged to highlight its position feature. The position similarity degree of a reference point with regard to the measured vehicle position is proposed and treated as weight to calculate vehicle coordinate, the low variability of CSI-fingerprint is alleviated. A filter is designed to cut off the dataset of remote reference points in localization algorithm, traffic noises are reduced. Finally, the accuracy and performance are verified in enforcement, autopilot and navigation applications. The effects of grid size, speed, traffic and weather are evaluated.
Vehicular Localization Based on CSI-Fingerprint and Vector Match
IEEE Transactions on Intelligent Transportation Systems ; 22 , 12 ; 7736-7746
2021-12-01
2311239 byte
Article (Journal)
Electronic Resource
English
Channel Covariance based Fingerprint Localization
IEEE | 2024
|Vector-Based Semantic Scenario Search for Vehicular Traffic
TIBKAT | 2024
|