Most train control systems require a human driver to interpret real time information and respond to this information. The human driver controls the vehicle following the defined rules. Adherence to the rules is often supervised with automated interventions to mitigate the risk of human error, for example, if the driver exceeds their authority (distance/speed). The driver/attendant can also react to unexpected events (degraded, abnormal and emergency situations).

The notable exceptions are the increasing number of automated metros (GoA4) that successfully provide high performance services within segregated environments with minimal variance (pre-programmed). The next evolution is to provide similar levels of automation and performance on open networks (mainline and light rail) where the environment cannot be easily controlled. Ultimately, as we move away from human driven vehicles and network management, outside the relative simplicity of the metro environment, autonomous systems will be required to handle all the unexpected events on the rail network, similarly to autonomous cars, necessitating the introduction of Artificial Intelligence (AI). Autonomous rail has less variables, but potentially more catastrophic consequences. The train control systems that are developed will need to consider the sensors available on-board and the information available from the connected network and how this information is processed and acted upon.

When managed correctly AI has the potential to significantly improve safety. However, to achieve societal acceptance research from the autonomous car industry suggest that this may need to be about two orders of magnitude better that the average driver. A suitable approach is required with society with respect to the use of AI. A period of increasing AI support for drivers (semi-autonomous), with a transition over to AI driving (autonomous), will be required. The expected approach is longitudinal testing, required for developing an inductive long-term dynamic safety case (‘proven in use’). Through aggressive training environments that use training sets and test conditions that place the system in ‘extreme’ operating environments and in-service experience with a driver present to ensure AI algorithms are able to operate safety with the eventual operational demands.

About the speakers:

David Milburn, Senior Technical Director, Global Leader – Digital Train Control, GHD, Australia

David Milburn is a Chartered Engineer with over twenty-years international railway experience.  David started his railway career when he joined the British Rail S&T graduate scheme. He quickly developed a keen interest in innovation and new technologies and in 1998 he was seconded to the ERTMS Users Group in Brussels, and has been involved in digital train control ever since.  David has held a number of senior roles, including leading the Signalling discipline for Network Rail’s flagship Crossrail Programme, and as the Technical Head, Signalling & Control Systems for Network Rail Consulting. Since moving to Australia, David has had the privilege of being involved in some fantastic projects, including: Sydney Metro Northwest, and a ground breaking ETCS Level 2 mine railway in WA. David is currently leading and developing the digital train control services for GHD.

Mike Erskine, Executive Advisor, GHD Advisory, Australia

Michael Erskine has over 34 years’ experience in Process, Mining, Transport, Water, Food, Oil & Gas, Infrastructure and Chemical industries. His consulting experience includes risk management, project management, process design and troubleshooting, accident investigation and feasibility study leader. Michael has significant experience in developing safety case and risk management documentation for government regulatory evaluation and organisational needs.

Registration Closes: Oct 17 2019 – 11:00 am