When cars learn to talk: the future of driver interaction
The increasing levels of automation of modern vehicles call for new forms of man-machine interaction. Together with industrial partners, researchers at the Fraunhofer Institute have developed an approach which takes this interaction to a new, situation-dependent level using AI. By Ruben Vorwald
Language models and sensors for more comfort
Just imagine: while driving, the car informs you of difficult road conditions and announces that vehicle control must soon be handed back to the driver. Scenes like this could soon become reality, thanks to the KARLI research projects which are working on AI-assisted interaction in the vehicle of the future. KARLI is “driven” by a consortium of Fraunhofer Institutes, Audi, Ford, Continental and other partners. The objective is to adapt communication between humans and vehicles to the respective level of automation, using language models and sensors which observe the traffic and the interior of the vehicle.
Levels of automation and the challenges they pose
Vehicles go through various stages of automation, from “assisted” to “autonomous“. KARLI focuses on stages two to four, i.e. partly automated to highly automated. This means that there are different requirements depending on the level: sometimes drivers must remain attentive, at other times they can relax. The interaction must be flexibly adapted. One example: in highly automated mode the vehicle could warn drivers of changes in the weather and prepare them to take over control. This demonstrates the strength of AI which processes information in real time and provides situation-specific information.
Personalised messaging through AI
AI does not only make interaction in the vehicle safer, it also makes it more personal. The system gets to know the user‘s preferences and adapts the message individually. Sensors register passenger condition and thus enable situation-specific communication. The objective is to improve driving experience by transferring vehicle control to the right person at the right time.
Potential for public transport: the tram of the future
Public transport could also profit from the findings of the KARLI project. Trams which increasingly use automated systems could adopt AI-assisted communication technology in order to increase passenger safety and comfort. Particularly in urban areas where trams play an important part in daily commuter traffic, it is vital that the systems function optimally and adapt to passenger requirements.
As in the car, personalized messages for passengers could also be made in trams. Warnings of road bends or braking procedures could help people to stabilize their position in good time. By monitoring passenger activities, AI could help minimise disruptions and enhance the overall travel experience. This way, the use of AI systems could help make urban public transport safer and more comfortable by optimising the interaction between traffic systems and passengers.