Rail must be bolder in its journey to automation

Rail must be bolder in its journey to automation
Dubai Metro
19 October 2020

By Tom Jansen, Global Domain Leader, Connected & Automated Vehicles, Utrecht NL

What has been holding back the wider adoption of automation in the rail industry? And what could be learnt from the advances in driverless technology made by the automotive sector?

These are the questions my colleague and I sought to answer when invited to take part in the IRSE series of Presidential lectures this month (see link at the foot of this page).

For me, the railway has always seemed the obvious home in which to nurture autonomous technologies.

After all, its promise is obvious: sophisticated technologies orchestrating train movements with greater precision; shorter headways and ‘emotion free’ decision making, coupled with reduced operator workloads and lower operational costs.

And it is not a recent concept for our industry. Early advances saw London’s Victoria Line open as one of the first applications of an automatic railway back in 1969.

Yet the International Association of Public Transport lists only around 60 fully-automated metro lines currently in operation around the world. For a technology more than half a century old, this represents surprisingly limited growth.

Beyond the world of metros, and some impressive innovations in long distance freight, something has been holding back the widespread adoption of autonomous technology across the world’s railways.

In our paper, we ask whether perhaps there is a more difficult underlying question: is the culture of the railway perhaps a little too safety oriented?

Automotive undaunted as it forges ahead

Where rail has shown restraint, the automotive sector has persisted with near unlimited budgets. Though many challenges must still be overcome - and a sense of realism is slowly setting in - the industry remains undaunted.

Vast databases are being built by big tech and automotive companies to calculate the endless scenarios that cars must manage on the roads.

But the industry is not waiting until vehicles can see around every corner of the world, or until every conceivable weather condition at every intersection at every possible time of day has been accounted for. Instead, the mindset is to forge ahead and accumulate knowledge along the way.

Drip feed innovation

Of course, the two industries are not on the same journey. For automotive, the emphasis is on customer experience. There is an expectation that one day it will be the market’s preference for vehicles themselves to take on the more mundane aspects of driving.

In rail, freeing up passenger time is not relevant. Instead, the benefits are about capacity and efficiency. Which puts safety under the spotlight.


Clearly, there are valid concerns about getting the ‘level’ of automation right or over the gradual erosion of key skills and competencies. Or indeed, of failures in the technology itself that could result in the loss of public confidence.

Safety is our industry’s highest priority. But this has embedded a highly risk-averse culture throughout the sector, the result being that whenever setbacks occur the industry’s instinct is always to retreat. 

For example, issues that emerge during early technology trials have seen entire programmes indefinitely postponed. Incidents on the tracks can lead to speed restrictions that degrade services far beyond the area in question, with the unintended consequence of pushing passengers away to other (less safe) modes.

Within such a cautious environment, innovation is restrained to a drip feed.

Fail fast and learn quickly 

In our paper we argue that the rail sector must be bolder about new ideas.

We highlight the Rivium ParkShuttle, a driverless service in Rotterdam that has been transporting approximately 2,500 passengers per day from a metro station to a nearby business park. Now more than 20 years old, the original pilot scheme thrived due to public and private sector collaboration that ultimately led to the ‘Experimenteerwet’ legislation enabling applicants to pursue individual implementations of autonomous technology on public roads in the Netherlands. As a result, subsequent projects gain opportunities to learn on a small scale, and to ‘fail fast’ before implementing the lessons in later developments.

But this approach is much less common in the rail sector, where it seems legislation needs to be fully developed beforehand.

We must ask ourselves: what happens if we fail to innovate? Will society be better off if we allow ourselves to become hostages of our own culture?

If appropriately applied, automation can deliver significant improvements in efficiency while simultaneously further enhancing our safety standards. However, it is an illusion to believe that we can know all variables of automation beforehand.

Let's be bolder. Let's take careful but meaningful steps whilst learning from each other and similar industries. Let's work to make the railways more competitive, efficient and affordable, so that our mode of transportation will still matter long into the future.

Tom Jansen
[email protected]

Tom's full paper will be published in the November IRSE News issue.
For more details: https://www.irse.org/Get-Involved/Events/Event-Details/eventDateId/1176