Abstract

Automatic Equipment Identification (AEI) tags are installed on all rail cars in North America to tag rolling stock and facilitate fault detection using wayside detectors. And yet, railcars do get lost when parked on sidings, etc., resulting in so-called Dark Cars. No effective solution has been proposed for finding such lost rail cars due to the distance limitations in the current AEI tag readers. However, recent developments in the area of AEI tags in particular multi-antenna wireless communication systems and backscatter communications, now provide the opportunity to read AEI tags from much farther distances. The adavntage of using multiple antennas is to increase read distance and directionality through beamforming while adhering to FCC power limits. The key obstacle to long-distance AEI tag reading is the limited link budget, which is the combination of the FCC-mandated power limit coupled with the minimum required EM field strength needed to power the AEI tag during a read operation. Since it is not feasible to modify the AEI tags themselves because a large number of them installed on railcars, nor boost the transmit power, our design is focused on increasing directionality of the transmitted power. In this work, the authors thus propose a novel use of a Multi-Antenna Beamforming AEI tag reader mounted on — and powered by — a locomotive. This is to detect AEI tags on dark cars by a train passing the dark cars. The authors conducted an in-depth model-driven performance evaluation of this scheme. The analysis includes different channel models, signal reflection, and impact from velocity on read success rates. The channel models considered in this evaluation include multipath models such as Rayleigh, Rician, and Two-Ray Ground. The results presented by the authors show that our proposed novel approach is not only feasible to read AEI tags in this fashion over distances exceeding multiple track separation distances, but also practical, in most cases requiring 10 or less antennas with only simple beamforming methods employed.

This content is only available via PDF.
You do not currently have access to this content.