For many years, traffic models have given us guidance in analyzing mobility problems and testing policy directions. But we live in a time where our traffic system is changing very quickly and in a flash, especially in the city. What does this mean for our traffic models? How do we ensure that they change sufficiently and can cope with the new and future reality?
The scope and nature of urban mobility is changing at a rapid pace. Major trends like increasing urbanization, increasing diversification in mobility behavior, always have access to information and the rapid emergence of new modes of transport and services such as e-bikes, self-driving vehicles and Mobility-as-a-Service (MaaS).
The latter two trends alone offer enormous opportunities – there is no shortage of dreams and wishful thinking in this field. But from a very sober point of view, they are all potential threats. An example: thanks to all those smart MaaS solutions, car ownership will soon fall (less parking space needed, more space for green space) or MaaS will mainly empty public transport (even more busy on the road). Staying ahead of developments and proactively adjusting traffic policy is therefore more necessary than ever because the importance of good governance is only increasing.
Traffic models are essential for such proactive good governance. As a (decentralized) government, you want to be able to anticipate, respond quickly and flexibly to new developments and also to accurately assess the effect of one measure aimed at one modality on another. Without the intelligence and computing power of models, that is practically impossible.
The problem is that the aforementioned trends are throwing our traffic system upside down in such a way that we can no longer cope with the traffic models that we have used over the past decades. In simple terms, they were primarily aimed at extending what we already know. What if the population grew by so many percent? What if we build a new road here? Or widen the road there? But we now need models that also help us explore the new and the unknown , such as the effects of MaaS and self-driving vehicles. What does that require from a model? Which model innovations are being worked on? And what could be smarter and better now?
Traffic models of tomorrow
If we want to be able to grasp and adjust developments in the mobility landscape somewhat, we will have to invest in a whole new generation of traffic models. That requires a lot from software developers and universities, but just as much from governments.
Suitable for many (types of) data
The mobility revolution is accompanied by a data revolution: millions of devices generate terabytes of data. Think of mobile phones: every time a device addresses a different mast, that device generates data. The movement data is ideal for calibrating your traffic model, but then you need software that turns the data into useful information and processes it into a traditional traffic model. The same applies to the data generated by connected vehicles and drones.
Today’s software can already process much of this data – see the Stravem article later in this issue – but the differences between software platforms are large and will become even greater. Moreover, every revolution has precursors who embrace the new possibilities and those who lag behind who continue to hope for better times.
We have known the concept of a circular economy for some time, but we must also focus on reuse in the area of data: we must move towards a circular (eco) system for traffic data and information. Many governments and companies notice that the data they generate and use in their work is also useful for other parties – sometimes parties from a completely different domain. Many governments have therefore set up open data platforms, often requesting (and rightly so!) To bring enriched data back to the platform: data-for-data contracts. At companies, the chances of reuse often lead to a data selling department that tries to sell data and information. In the world of traffic models, circularity has now become commonplace. Whereas in the past a large part of the project budget was spent collecting project-specific data such as counts, household surveys and measurements, the start of a modeling project now mainly consists of combining data from dozens or perhaps hundreds of different sources. This has the additional effect that the same data is used in multiple projects, which ensures greater consistency between projects. That can only be welcomed.
One software environment
A third development or requirement: we see a move towards a single software environment that can connect all studies on an operational, tactical and strategic level, feeds back and forth with information and allows them to interact with each other, and thus influence them.
That is really becoming a necessity, if only because of consistency. After all, traffic policy can only work if it is consistent across all levels. An example is the impact of the sharing economy on our traffic. More strategic studies such as the Lisbon Study and the follow-up study in Helsinki show what the long-term consequences could be – and from those studies we can dream of more than a halving of traffic to the release of 210 football fields of space in the city. But how will things work out in practice? In a follow-up study, the ITF simulated the possible impact of pick-up & drop-off zones on traffic flow. These operational studies show that even with a low penetration rate of ride sharinga serious disruption of the flow can occur, depending on how picking up and driving off is organized. This process is exemplary of studying the consequences of a social trend at all levels, from operational level to strategic level, and from short to very long term. The different traffic models used at these levels must really be mutually consistent.
Software providers are increasingly moving towards switched software solutions where it is easy to switch between the layers without data loss.
Predicting the impact of new technology
A fourth point that we want to mention concerns the new technology that is approaching us: a traffic model must be able to predict the impact thereof. No matter how simple this is worded, it is by no means simple. The technology itself continues to change and is therefore not a given. But also the user acceptance (is the technology being picked up?) Is a tricky one. It just depends on who you ask: one person thinks we will all be driving an autonomous car within five years, the other one expects the horse and carriage to return. The truth will lie somewhere in the middle, but what we demand from traffic models is to provide insight into those many possible “futures.”
Many model scenarios are therefore no longer built to calculate the most likely future scenario , but rather to calculate extreme scenarios – to explore the impact of a scenario that may never happen. Even the largest non-believer needs to study the impact of autonomous and shared motoring, only to be able to demonstrate why this will never happen. A public transport plan that does not take into account the role of on-demand shuttles in future public transport is not a future-proof public transport plan, and a station design that does not provide space for sharing bicycles is obsolete before it is built. The traffic models must be able to demonstrate how it could function.
As a fifth and final point, we would like to note that the market increasingly demands a holistic approach. For many traffic-related issues, we are not only interested in traffic itself, but also in aspects such as air quality, noise emissions, electricity and economic effects. Electricity is a logical component when you consider that the arrival of autonomous vehicles may go hand in hand with electrification and that these vehicles cannot all be charged at the same time. Regarding the economy: a decrease in car ownership will automatically result in the loss of parking spaces, resulting in loss of income. We will also order fewer cars, as a result of which the revenue model changes from paying for possession to paying per use. This can in turn lead to companies acting according to the slogan ‘to seduce the customer to make as many kilometers as possible’, in order to remain profitable. The government must also manage these economic effects and continue to guarantee safety, quality of life and accessibility.
Traffic models and simulation tools are becoming increasingly important to keep a grip on the traffic system and everything related to it. The traffic model software has evolved considerably in recent years to be ready for tomorrow’s questions. Our mobility system can go in different directions, depending on the choice that policy makers make today. It is therefore more important than ever to use the right tools to answer these questions. But beware: as with every revolution, there are leaders and laggards. So keep looking critically at the software that is used, because without good model software, good mobility policy is difficult.