A Railway Accident Prevention System Using an Intelligent Pilot Vehicle

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

1 Scopus Citations
View graph of relations

Author(s)

  • Shixiong Wang
  • Xinke Li
  • Zhirui Chen
  • Yang Liu

Detail(s)

Original languageEnglish
Pages (from-to)5170-5188
Journal / PublicationIEEE Transactions on Intelligent Transportation Systems
Volume25
Issue number6
Online published1 Dec 2023
Publication statusPublished - Jun 2024
Externally publishedYes

Abstract

Railway transportation, as a pillar of modern civilization, unavoidably suffers from external risk factors such as natural disasters, track breakages, and train collisions, which lead to substantial loss of life and property. Therefore, there is an urgent need to design a mechanism for warning and preventing railway accidents in order to diminish costs. We propose an add-on solution to the current system, which equips a train with a multifunctional pilot vehicle in the front: the vehicle pilots its mother train, warning it of impending danger, and stopping it if required. Specifically, the pilot vehicle is equipped with a wireless communication device to converse with the mother train, a ranging device for measuring the real-time distance from the mother train, a camera to capture the railway conditions ahead and recognize anomaly situations, and other sensors (e.g., collision detector and tiltmeter) to monitor its own conditions. Based on the above equipment, an efficient autonomous driving method is designed for the pilot vehicle to adjust the distance from the train. The autonomous driving problem can be formulated into a multi-objective functional optimization, where the objective is to minimize the total energy consumption and the experienced jerk of the pilot vehicle, and the decision is a continuous-time function that represents the traction or braking force imposed on the pilot vehicle. Additionally, a vision-based deep learning method is devised to automatically detect the mentioned railway anomalies using the ego-view camera of the pilot vehicle. To control the operational and maintenance costs, we propose to deploy pilot vehicles only for trains running in potentially dangerous environments, e.g., mountainous areas during rainy days. By implementing the proposed scheme, we anticipate a reduction in accident rates within railway systems. © 2023 IEEE.

Research Area(s)

  • anomaly detection, autonomous driving and control, deep neural network, Intelligent transportation, multi-objective functional optimization, railway accidents

Citation Format(s)

A Railway Accident Prevention System Using an Intelligent Pilot Vehicle. / Wang, Shixiong; Li, Xinke; Chen, Zhirui et al.
In: IEEE Transactions on Intelligent Transportation Systems, Vol. 25, No. 6, 06.2024, p. 5170-5188.

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review