West Midlands Trains:
A new era in competence management
'We’re using information from on-train data recorders in promising new ways', says Adrian Champion, Head of Operational Standards at West Midlands Trains.
For Adrian Champion, Head of Operational Standards at West Midlands Trains, automated driver competence indicators hold a lot of promise. Among other things, they supply the operator with incredibly detailed information. Here’s how, in his words, they can make a powerful difference to driver training and competence management.
Adrian ChampionHead of Operational Standards, West Midlands Trains
What do on-train data recorders do?
On-train data recorders (OTDRs) record information about the operational controls of a train and how the train performs when those controls are operated. They’re similar to black boxes on aircraft.
OTDRs aren’t new technology. We’ve been using them as a method of assessing drivers and doing investigations into train performance for many years.
But with the help of a piece of software called TrainServ, we now have the ability to analyse and compare this data over longer periods of time and across multiple trains, rather than just looking at one train on one journey. This gives us the ability to automate driver competence indicators.
Why did West Midlands Trains trial OTDRs in 2023/24?
The industry’s approach to competence management has been relatively static for the past 30 years and largely involves a manual process of gathering data. But I’ve always been interested in finding the next era of competence management.
When I saw RSSB’s research project on automated driver competency indicators (ADCIs), I knew the TrainServ software would provide competence evidence over a longer period of time, which would give us an opportunity to better understand consistency in human performance.
Another benefit is the ability to see more positive indicators of driving performance. That interests me as well. These positive indicators highlight consistent, skilled, competent drivers, which demonstrates how infrequently incidents occur when compared with driving hours and helps us better understand how to prevent incidents. Current methods largely rely on the absence of incidents as a positive indication, which is a weak form of evidence.
Through their strategic partnership, RSSB and the University of Huddersfield tested the validity and value of the insights that could be gained from automated driver competence indicators.
The industry invests significant time and effort in assessing train driver competence, including a good deal of manual data analysis. The research team wanted to understand whether automating the processing of OTDR data could save time and provide more granular information.
Software developer Cogitare, with its specialism in rail, adapted its TrainServ software to integrate the OTDR data with infrastructure, signalling, and timetabling data, allowing ADCIs to be generated.
For West Midlands Trains, which participated in the 2023/24 trial, there’s certainly value in using this method to enhance competence management.
What did you want to investigate during your trial?
We wanted to understand whether we could gain assessment or competence evidence differently.
In particular, we wanted to enhance the way we train our drivers, as well as the way our drivers understand error management. Being armed with so much more evidence—including the positive evidence I mentioned—allows us to balance our understanding of how errors can occur. We can feed that information back into our training and management programmes.
Once we began the trial, we realised there was a lot more potential. We knew there was more that we could invest back into the operational railway. We’re only one operator, with a relatively small amount of data. Imagine what could happen if we scale this up to lots of different operators.
What did the trial reveal about human factors?
This trial has reminded us that we should be expecting our drivers to make mistakes. They’re human. Traditionally, our response to an incident or when someone’s made a mistake is to look at that particular journey on that particular day and try to understand what caused it. But as I said earlier, we can now use the positive indicators—across the hundreds of journeys a driver has made—to balance our understanding of what happened.
If a driver has made an error, say, once in 300 journeys, we have a lot of positive evidence of skilful driving to use as a balance, rather than focussing on any incident that may have occurred as a result. We can help the driver process what’s happened, and we can learn how system changes can help prevent incidents and even start to predict when human error or system failures are more likely to occur.
Having access to data over longer periods has been eye-opening and will allow us to look at incidents, human performance, and competence management with a different perspective—one that’s more in line with the fair culture we strive for.
How might the trial change your approach to competence management?
Because the positive indicators have revealed that we have a lot of skilled drivers, we can move away from focusing on an absence of incidents or taking a microscopic approach following an adverse event.
We don’t want to be saying to drivers, ‘You need to change because you’re not driving in line with the policies’. Instead, we want to recognise that they’re the experts, and we should be learning from them to inform our future training programmes, policies, and driving styles.
We’d also like to give drivers access to this information directly so that they can see for themselves how they’re driving and where they can make changes, if necessary. Everybody wants to do the best they can, and we’d like to give them the opportunity to adjust their style themselves.
We’re already looking at how we can use these principles to develop our existing processes to foster a more transparent safety culture.
Lastly, and more broadly, why is using data in this way important to West Midlands Trains?
Having access to data covering many train journeys over a long period of time provides reassurance and context that, most of the time, everything is working as expected—rather than microscopically reviewing small amounts of data when things do go wrong.
For example, we can look at how different times of the day or year might affect driving styles, or what we can learn when comparing driving in periods of low adhesion with driving in the summer months. We can then feed this information into timetable planning and infrastructure maintenance, as well as learn more about human performance.
Data like this can help safety systems evolve, feed into training and development programmes, and improve overall system risk. That’s quite powerful.
Why not check out the research that started it all? Head to our Research Catalogue for more.
Take me there ❯
