One of the factors contributing to transport-related emissions is driving behaviour. In this third interview of the MODALES project, the Luxembourg Institute of Science and Technology (LIST) explains how driving behaviour will be measured within the project.

LIST, what is your role in MODALES?

As Research and Technology Organisation, we are in charge of proposing and validating new technological approaches to encourage low-emission driving practices. In 2021, LIST is for instance going to create a mobile assistant app for low-emission driving, with the aim of supporting the project’s experimentation activities. Our team working in MODALES is involved in research activities related to ubiquitous and mobile computing, but also to network performance evaluation, which are two research areas that are closely related to the project. MODALES offers an interesting perspective to these domains as the objective is to offer active driving analysis solutions in areas that are sometimes poorly covered by cellular networks and with several mobility constraints. This requires new approaches to be exploited. Beyond this responsibility, we also coordinate the work on creating guidelines for low-emission training, also in collaboration with ACASA and IRU, and develop related tools, such as a web dashboard with CERTH. LIST is also involved transversally in other project tasks, either theoretical (e.g., study of correlation models, some of which will be integrated in the mobile app) or experimental (e.g., data collection campaigns, which will mostly rely on the mobile app).

Within MODALES, LIST will develop a mobile application for users to better track their driving behaviour and assist them in low-emission driving. Tell us more.

Today’s phones come with a tremendous number of sensors and computing capabilities, allowing them to accommodate innovative applications and interactions with the user. In MODALES, we intend to exploit these capabilities by creating a real-time mobile intelligence to understand how a driver behaves, and subsequently give recommendations for low-emission driving. To do this, we are currently designing two modules.

The first one collects the key data that will serve as the basis for understanding a user’s driving behaviour. To that aim, we rely primarily on the sensors that are integrated in all modern phones, like accelerometers and gyroscopes. Under certain conditions, these sensors can be used to deduce the accelerations and decelerations patterns experienced by a user’s vehicle, and thus interpret various behavioural patterns. Other data, such as anonymous wireless network information, are also considered to estimate contextual indicators that have an influence on low-emission driving, like the presence of a traffic jam or environmental/road conditions. In addition to phone sensors, we also plan on using On-Board Diagnostic (OBD) dongles, which can be connected to a standard port available in most vehicles, to collect data directly from the vehicle (e.g., RPM – revolution per minute), as well as open web services and data sets (e.g., weather database).

The second module aims at using these indicators to generate recommendations for the driver. In MODALES, recommendations will be given both while the user is driving and after his/her trip, in the form of a detailed report. They will consider the impact that user behaviour has on the emissions generated by exhaust gases, brakes and tyres. This mechanism induces fascinating research challenges, but also the need for attractive and non-intrusive user interfaces. For this we are considering approaches like gamification, which consists in using game mechanisms in the application to encourage the user to keep using the app on a regular basis.

Before developing these two modules though, we need to identify and establish a precise list of factors that influence driving emissions, whether these are related to the environment (e.g., weather, traffic congestion), the vehicle (e.g., fuel, age) or the driver (e.g., driving experience), so that adequate recommendations can be generated accordingly. It is important to mention at this stage that the project team will do its utmost to develop a mobile intelligence that minimises the use of an Internet connection as much as possible, and which would therefore be able to operate independently. Data that will have to be stored on a server will respect strict rules of anonymisation as established by the GDPR.

When will this app be available? How many people do you hope will be using this app?

The development of the mobile app will be done in several phases. We plan on testing and validating different versions in 2021, in parallel with the setting up of experimentation sites. We will start in the first quarter of 2021 by testing the data collection module in Luxembourg and Barcelona. This will serve as a basis for the implementation of driver profiling models and the recommendation module. The final app should be ready in the third quarter of 2021, so that it can be used as the main measurement tool for the experimental sites of the project. In total, we aim to reach several hundred users across eight experimental sites, ranging from private and professional drivers and several types of vehicles, from cars to trucks and buses.

How important is research in this kind of project and why?

Research is fundamental to improve our understanding of low-emission behaviours, and to implement strategies to reduce them. The research topics covered in MODALES are multidisciplinary and involve partners with a variety of expertise. These skills involve, for example, understanding the variability of emissions from exhaust gases, tyres and brakes according to a certain behaviour and context, but also interpreting indicators using artificial intelligence techniques, or creating precise mathematical models. All this is an absolutely necessary basis for achieving the objectives of the project.

LIST is based in Luxembourg. What do you hope your contribution to MODALES will bring to your country?

Luxembourg, with 662 passenger cars per thousand inhabitants, is the EU Member State with the highest motorisation rate relative to population (Eurostat data for 2016). Furthermore, in normal times more than 200,000 commuters come to Luxembourg every day from the bordering countries, with huge impacts on congestion and air pollution, especially in the peak hours, which are aggravated by sudden acceleration and braking patterns. Solving this problem will certainly require structural measures such as infrastructural investments to encourage modal shift and the increasing electrification of the fleet, but these will be effective in the longer term. In a shorter time, MODALES could offer a tangible and swift solution to improve driving behaviours and vehicle maintenance, reducing the noxious emissions associated with the current traffic.