Bus driver and Assistant Professor: a unique double position

After a failed Mathematics degree programme, Marcel van Kooten Niekerk joined in public transport in the 1990s. And even though his academic career took off after all in the years that followed - he works at the AI & Mobility Lab on possible applications for artificial intelligence in public transport, he still loves to get in the driver's seat of the public bus.

A failed student. That is what Marcel van Kooten Niekerk was when he obtained his bus driver's license for fun in the late 1990s. He had just quit his degree programme in Mathematics and did not have a clear plan for the future yet. “Public transport always interested me. During my degree programme, I also developed a travel planner for public transport, a primitive version of the apps we currently have on our phones.”

Bus company Connexxion showed an interest in his planner. Van Kooten Niekerk got a job there to further develop the travel planner, but also joined as a bus driver AND made schedules for bus rides and staff. “I liked the variety – three days at the office and two days on the bus, but I also found it important. I saw how my solutions worked out. Some solutions work perfect in theory, but so many more factors matter in practice.” He later moved to Qbuzz, where he became a Logistics and Innovation advisor.

Doing research yourself

Twenty years later, he still works on logistics and innovation, but as an academic, at the AI & Mobility Lab. This is one of the fifteen AI Labs of Utrecht University, in which the university collaborates with experts from practice on AI applications to societal issues. An example of this is that in the AI & Mobility Lab, researchers from the university collaborate with specialists from, among others, Qbuzz, ProRail and NS on innovative AI techniques to make public transport more robust and future proof. Van Kooten Niekerk supervises students in their graduation and PhD research since May 2023.

How does a planner/bus driver find his way back to the classroom? Van Kooten Niekerk says: “In 2006, I was at a conference in Leeds, where I met Marjan van den Akker. She was an assistant professor at Computer Science here in Utrecht at the time and she wanted to know whether or not I could supervise students in their graduation research. I became so enthusiastic about it that my wife said: ‘Wouldn't you much rather do research yourself?’ She encouraged me to study again.”

Some solutions work perfect in theory, but so many more factors matter in practice.

Van Kooten Niekerk enrolled and obtained his Bachelor’s degree in little more than a year, besides a full-time job. This was followed by a Master's degree in 2010 and then a PhD track, with Marjan van den Akker – an associate professor at that point – as one of his co promoters. All of this besides his job at Qbuzz. “That was quite much, but the good thing was that I could use everything I did for my education and research at my job too.” For his PhD defence, one of the things he developed was software for scheduling rides for electric buses, which were just being introduced at the time.

“An electric fleet requires a different schedule, because you have to consider limited battery capacity and charging times. The bus can't ride for an entire day without charging the batteries. By now, about 30 – 40 percent of the fleet is electric and the goal is to only have electric buses on the road in 2030, so this schedule will only become more and more complex.”

AI and the energy transition

He is currently working on a project in the field of energy transition. With the help of artificial intelligence, Van Kooten Niekerk and his students' algorithms calculate what is a convenient time to charge. “Imagine a bus has a delay of ten minutes, but twenty minutes have been scheduled for charging time. Our software can then calculate very quickly and based on previous experiences what the best solution is: charge for ten minutes to sort out the schedule, charge for fifteen minutes and accept a delay of five minutes or not charge at all and have the charging moment take place at a later time, like at the end of a shift.”

That is essentially no different from the software Van Kooten Niekerk developed for his PhD research, with the biggest difference being that the processing power has since become much bigger. “Because of this, it now takes two hours instead of two weeks to make a schedule. That provides space for schedulers to look at variants and to see what works out best for the employer AND the employee. For instance, each driver is only allowed to drive four hours back to back. You can have a schedule made in which everyone neatly drives those four hours. But with help from AI, you can sometimes also make another schedule which works just as well, does not cost more, but means in practice that a driver drives for 3.5 hours. So you provide a more pleasant workday for the bus driver.”

Another issue AI can help in is to charge as sustainable as possible. “Charging the batteries of a bus is a big strain on the power grid. And the demand is already high to begin with. With help from AI, you can calculate the most convenient moment to charge and then make agreements with various parties about who uses the power grid at what time.”

If I were only a researcher in the field of public transport, I never would’ve had access to the knowledge I have now.

More sustainable and more efficient

Artificial intelligence will not make human schedulers' work obsolete, Van Kooten Niekerk believes. “Thanks to the use of AI, we can make better informed decisions. The fleet is used more sustainably and more efficiently, but the staff schedules and the detection of technological malfunctions in the bus become better too. For human schedulers, keeping an overview of this becomes more and more complex and AI can assist us in that.”

At Qbuzz, Van Kooten Niekerk still does drive buses every once in a while. “Not that often, though. I think about once a month, such as if there is a hole in the schedule. I can very easily drive for two hours and then come to the university. The contact with people, providing the service, that's what I enjoy.” For now, Van Kooten Niekerk will keep up this double job for a while. “This combination is quite unique. The entire chain, from research to scheduling and driving, I do myself. If I were only a researcher in the field of public transport, I never would’ve had access to the knowledge I currently have. That really has added value, for all parties.”