Development of visual perception systems for autonomous trains using deep convolutional neural networks

This project aims to develop a deep learning-based computer vision system of a high-performance visual perception system to support the Autonomous Rail. The developed visual perception system enables extracting rail tracks and objects in typical rail scenes using consumer grade camera and lidar sensors mounted on a moving rail. This visual perception capacity is fundamental to understand a self-ego motion and surrounding environment for supporting the Autonomous Vehicle Recovery (AVR) capability of Level 1 (L1) Autonomous Rail. With the AVR capability, the vehicle moving with a low-speed can make decision and acting as a human operator and successfully drive itself to the next stopping point. For achieving goal, two intern students at York University will closely work with a multidisciplinary engineering team at Thales Canada during eight months. The research outcomes obtained through this project will be evaluated with extensive benchmark data generated by the research team.

Faculty Supervisor:

Gunho Sohn

Student:

Partner:

Thales Canada Inc (North York, ON)

Discipline:

Computer science

Sector:

Professional, scientific and technical services; Transportation and warehousing

University:

York University

Program: