Professor Serge Hoogendoorn is Professor of Traffic Management at the Faculty of Civil Engineering and Geosciences (CEG). In 2015 he received an ERC Advanced Grant for five-year research into traffic theory for pedestrians and cyclists. ‘Pedestrian and cyclist traffic is becoming ever more important in our cities. In Amsterdam it is already the most important mode of transport,’ according to Hoogendoorn.
As a 17-year-old, I thought of studying science, as I was good at maths and physics. It was either that or visual arts. The academy of arts attracted me immensely, but in the end I decided that studying something with better financial prospects was the wiser option. After a visit to the TU Delft open days my choice was easy: applied mathematics.
It suited me very well. I was most fascinated by the subjects that had more of an applied character. I found the system theory subjects truly inspirational. There you learned how to measure systems, capture them in mathematical terms, and then analyse and optimise them. I found that both an inspiration and a challenge. I chose my graduation subject after a course in optimum control of distributed parameter systems. After three lectures I had completely lost track of the subject and I couldn’t stand that. Following my graduation I was supposed to enter mandatory military service, but shortly before that, my graduation professor phoned: would I be interested in a PhD on modelling and optimising traffic streams? I was never conscripted in the end.
My field of research concerns the observing, understanding and modelling of traffic flows in networks. The aim is to use the resulting knowledge and models to improve the design, planning and flow of traffic. Improving the flow of traffic can be done in many ways, for example with assisted driving systems or by influencing traffic with the help of traffic lights or ramp metering systems. My main interest are the control mechanisms that allow you to optimise this. For example, I would like to understand the best way to divide the flow of traffic in a network in order to minimalise congestion. I don’t think the essential question here is whether you do that with the help of variable messaging signs along the road or with systems in the car, but I’ve noticed that this discussion of in-car vs. roadside is getting a lot of attention.
What we do is fundamental research from an applied perspective. This usually begins with the gathering of data, from practice or sometimes through driving simulators or questionnaires. These data provide insights into the behaviour of road users and these insights in turn lead to new theory and mathematical models. During my NWO-Vidi research project, we registered the driving behaviour of motorists on motorways from a helicopter. A thorough analysis of the data showed us how to capture this in accurate models.
The models can then be used to predict for example what will happen if the share of lorry traffic increases by twenty percent or if we lower or increase the maximum driving speeds. Next, you can think of ways to optimise the traffic system. You can do this with simple scenarios such as an overtaking prohibition or an extra traffic lane, or with elegant mathematical optimisation techniques.
You want to test such optimisations in practice of course. The Praktijkproef Amsterdam (‘Amsterdam Practical Trial’) is a good example. It is the first large- scale pilot with coordinated network traffic management in the world. Commissioned by Rijkswaterstaat, we are collaborating in this project with the Ministry of Infrastructure and the Environment, the municipality of Amsterdam, Province of Noord-Holland, City Region of Amsterdam, and various companies. The aim is to improve the traffic situation in and around Amsterdam. This can be achieved by attuning the available measures such as traffic ways and ramp metering as best you can. Done sensibly, it will achieve a definite improvement in the flow of traffic in the network.
The trial does not only help decrease traffic jams, but also provides us with many new insights. It turns out that some of the accepted ideas on the control of traffic are not as important as previously thought. For example, predicting the exact moment a traffic jam will arise is not only impossible, but it is not really necessary here either. It is better to react effectively than to anticipate wrongly. The exact location and length of tailbacks are much more important for coordinated network traffic management, and so is information about the routes that motorists choose. In this practical case, the question is how we can get a clearer view of actual and expected traffic situations. Possibly so-called floating car data, i.e. information from navigation systems and smartphones in cars, can help here.
My ERC Grant means recognition for the quality of our work, and the funds enable us to perform more exciting and pioneering new research. I believe that slow traffic as a research subject will only gain importance over the coming years. Not just from a scientific standpoint, but also because walking and cycling are more often the preferred modes of transport in overcrowded cities. This is already the case in Amsterdam, and it has clear advantages for accessibility, liveability and health. Hence, we should take care that cycling and walking in the city remains attractive. Bicycle traffic jams are a hot topic, and problems with bicycle parking near train stations are regularly on the news, but the interaction of bicycles and pedestrians with other modes of transportation can also lead to problems with safety and traffic flow. Not just right now, but also in the future when self-driving cars will be making their entrance. We need more knowledge to cope with these problems and to come up with smart solutions for them.
‘We have become one of the most important research groups in the field of pedestrian flow theory and management’
Since the late 1990s we have been successfully pioneering the monitoring and control of pedestrian crowds. We have become one of the most important research groups in the field of pedestrian flow theory and management. In traffic engineering however, car traffic is still receiving more attention than pedestrian traffic, let alone cycle traffic. That has probably to do with the impact of cars and car traffic (and traffic jams) on people and society. I think of things like the economic cost of congestion and the frustration caused by it, or the freedom cars seem to offer people and the social status they derive from their cars. In view of the continuing urbanisation it is necessary to shift focus to urban mobility, including motor traffic and public transport, but also including pedestrians and cyclists. This will certainly be my focal point in the coming years.
A car is not always the most attractive mode of transport anyway: in many busy cities it is easier and cheaper to travel from A to B on foot or by bicycle. Moreover, as our cities, including our train stations and other public spaces, are becoming progressively busier, the available space for cyclists and pedestrians is getting scarcer too. The call to improve the use of our public spaces is getting louder; this to ensure continuing safety, accessibility and comfort. Thus cycling and walking are getting more and more attention. The car as a holy cow might lose some of its sanctity in the coming years. Owning a car may not be as important to the younger generation as it was twenty years ago.
We are ultimately concerned with the scientific challenge. This field of research is still in its infancy, as it were. There is a distinct lack of mathematical models and theory describing cyclists’ behaviour. We have taken good first steps in the past for pedestrian flows and crowds, but here too we are still far away from a generally accepted theory. It is not that surprising. On the one hand it is simply difficult, as it involves the behaviour of many individuals each with their own motivation, who – often subconsciously – communicate with each other, interact and move in all directions. On the other hand there are so few data available. Data are now becoming available in dribs and drabs, allowing us to develop methods for modelling the behaviour of cyclists and pedestrians. The ERC grant is a great opportunity to be at the forefront of this development.
We do not work in a vacuum. Mobility is a multidisciplinary field by definition. We have been collaborating with scientists from various domains for many years. My Vici project, for example, was aimed at modelling and managing traffic and transport under unusual circumstances and in disaster scenarios. The research team comprised psychologists, civil engineers, mathematicians, modelling experts and information scientists. This collaboration allowed us to make real progress, not only in understanding human behaviour in extreme situations, but also in formalising that knowledge into mathematical models. Behaviour is important, of course. Traffic jams or pedestrian flows along a busy shopping street; it is the eventual result of human behaviour. Human behaviour at different levels at that, from the split-second decision of braking for a pedestrian crossing the street, to the long-term decision you take when you consider cycling to work tomorrow instead of going by car.
This multidisciplinary collaboration can also be seen in the Amsterdam Institute for Advanced Metropolitan Solutions (AMS). As Principal Investigator I try, together with my AMS colleagues, to combine the urban mobility research with problems or solutions in other fields whenever possible. We view the city as a Living Lab, so we intend to test out things, allowing us to learn by doing, rather than trying to figure things out sitting behind our desks. We really need such a lab to develop robust solutions for an unruly reality. AMS is also important for the ERC research project. Our close relationship with the municipality and other partners enable us to gather data and conduct experiments in the city. This can involve a large project like the setting up of the Urban Mobility Lab, but also a smaller initiative like equipping AMS employees’ bicycles with GPS. Moreover, AMS is a group of very enthusiastic and driven people who I really enjoy working with.
For larger projects such as the Amsterdam Practical Trial it is vital to collaborate with businesses and governments. Working closely together with a company like INCONTROL Software Solutions also provides you with insight into what problems are encountered in practice and hence what is really needed. I often see scientists making assumptions about what is needed in practice without really getting in touch with people out there. That gap is too wide, and I honestly feel that this is not acceptable in a field like ours. It is for this reason that my group undertakes a lot of applied research in addition to fundamental research, and we are all engaged in disseminating knowledge to the professional sector. Besides, our intensive collaboration with businesses and governments also makes it easier for our students to find an external graduation project or an exciting internship.
Practical trial in Melbourne
Professor Serge Hoogendoorn and his team are not only crossing disciplinary borders in their research, but actual ones as well. ‘In Melbourne we are currently setting up a similar practical trial to the one in Amsterdam,’ says Hoogendoorn. ‘I’m doing this together with Professor Hai Le Vu of the Swinburne University of Technology in Melbourne, where I am an honorary professor. The issues are different in Melbourne, so I expect to learn a lot. Central to the problem is that the motorway network is not finished: the motorway becomes an urban connection road and then becomes a motorway again. This leads to problems, as traffic has no easy access to the city. We are now in the process of adapting our methods of measuring and controlling to this situation. This is bound to be quite a challenge, because local traffic data are of a lesser quality than in the Netherlands.’
TU Delft Highlights 2015
De Highlights are a document that reflects TU Delft and the highlights of that year. In the Highlights 2015 this interview with Serge Hoogendoorn was published.