Towards artificial intelligence based rail driving assistance tool

Abstract

This work proposes additional levels of progressive driver assistance expanding the traditional Grades Of Automation (GoA) in order to allow both higher level of automation and keeping the driver involved in driving task at the same time. The second contribution is the Digital Co-Driver which aims to bring the driver back in the train driving activity with the new GoA defined before, taking into account human involvement and driving skills. This framework is made up of several modules, each of which addresses a specific issue arising from the increased level of automation. The Driver State and Performance Monitoring Module monitors the driver’s involvement, situation awareness and performance. The Digital Adviser Module improves driver’s situational awareness, and the Digital Teacher Module improves his/her driving skills and knowledge of the system. Finally, the Safety Manager ensures the system’s compatibility with safety standards.