Artificial intelligence (AI) has already begun to fundamentally reshape the way we understand and use transport. From traffic prediction to autonomous driving, technology is moving from being an auxiliary tool to being a central decision-maker.
In Europe, research centres and universities are systematically investing in the development of Intelligent Transport Systems (ITS), with the aim of optimising mobility, enhancing road safety, reducing emissions and serving citizens through smart infrastructure.
From traffic prediction to understanding driving behaviour
Modern AI applications use data from sensors, mobile phones, cameras and GPS systems to train predictive models that can recognise traffic patterns, identify anomalies, calculate delays and suggest alternative routes in real time.
In parallel, algorithms are being developed that study human driving behavior, analyzing data such as speed, distance from other vehicles, braking patterns, and reactions to external stimuli. These models have a dual purpose: on the one hand, to enhance safety by identifying dangerous practices, and on the other hand, to feed autonomous systems with a “human” perception of the road.
In many cases, simulated data is used in virtual driving environments to train and test these models without the risks and uncertainties of the real world.
Autonomous driving: from perception to decision
The development of autonomous vehicles is perhaps the most ambitious bet in the sector. These vehicles rely on a combination of sensors (LiDAR, cameras, radar), perception techniques, and decision-making systems that allow them to “understand” their environment and act on it.
A key factor is the ability of these systems to react appropriately in complex, uncertain or even morally ambiguous situations — for example, in a dilemma where avoiding an obstacle may create a new risk.
In this context, methods based on risk assessment methods are adopted before each decision is made, such as Failure Mode and Effects Analysis (FMEA), an approach that allows the prioritization of possible errors and their effects, with the aim of enhancing the reliability of decision-making systems.
Foundation models and generalized intelligence for driving
In recent years, research has also turned to the so-called foundation models, i.e. large, general artificial intelligence models that can be adapted to many different tasks. These are models that have the ability to “learn” driving behavior, object recognition, movement prediction and navigation in a single, multimodal framework.
The integration of data from multiple sources (cameras, maps, CAN bus signals, motion sensors) allows these models to operate effectively even in complex, urban environments. At the same time, the use of “prompts” allows designers to guide the vehicle’s behavior — for example, choosing “conservative” or “aggressive” driving depending on the conditions.
The end-to-end approach, where the model receives sensory data and directly generates vehicle control commands, is also being tested with impressive results, although it is still considered premature for full adoption.
Safety, ethics and social responsibility
Beyond technical excellence, the development of AI systems for transport must take into account the social and ethical dimension. Who decides in the event of a collision? How can we ensure that technology serves all citizens equally, without discrimination or exclusion?
European research projects are developing equality-sensitive tools that examine how different social groups (e.g. women, the elderly, people with disabilities) are affected by technological interventions in public space and transport. At the same time, researchers are studying the impact of autonomous vehicles on the design of cities and the spatial distribution of infrastructure.
The connection of artificial intelligence with mental health is another emerging trend: embedded systems in vehicles or mobility platforms can detect signs of fatigue, stress or confusion, offering early intervention to protect the driver and passengers.
Research with a European footprint
Europe maintains a strong presence in research on intelligent transport, with a multitude of projects under Horizon Europe, the Chips Joint Undertaking (JU), the EIT Urban Mobility and other frameworks. Universities and research centres from the Member States are actively involved in this effort, developing cutting-edge technologies and collaborating with industry, local authorities and international bodies.
The focus on issues of transparency, ethics and social benefit gives the European research area a distinct identity compared to other regions of the world, where the market drives developments rapidly but often without sufficient public control.
Artificial intelligence in transport is no longer a futuristic vision. It is already present on the roads, in vehicles and in infrastructure. The big challenge is to shape it in a way that serves people, enhances security, and contributes to a more sustainable and accessible world.
