In a new research project, researchers from the University of Skövde together with Scania, Smart Eye, and Viscando are set to develop new AI systems for cars and trucks. These systems are expected to enhance road safety by improving interaction between drivers and AI. The project is funded with 7.5 million SEK from FFI-Vinnova and 7.6 million SEK from the industries.
Imagine yourself on the roads in your car, when suddenly, the weather turns severe. The rain intensifies, drastically reducing visibility. It becomes challenging to see road markings and other vehicles due to the heavy rain. This is where driver assistance systems can be invaluable for your safety and that of others, assuming you have these safety systems activated.
Paul Hemeren, Associate Professor of Informatics, leads the University of Skövde's work in the I2Connect project: Intelligent, Interactive and Connected: Next Generation of Realtime Driver Safety Support Systems. In this project, he and his colleagues will develop new AI safety systems, known in the industry as ADAS (Advanced Driver Assistance Systems), for truck and car drivers. The aim is to enhance safety by improving the interaction between drivers and safety systems.
Building trust in the systems
A functional interaction between drivers and safety systems is crucial for these smart systems to be utilized. We must be willing to trust that these smart systems are indeed intelligent. That's why the interface between humans and machines, known as HMI (Human-Machine Interface), is so crucial, according to Paul Hemeren.
"Currently, many drivers either deactivate the safety system or react negatively to the system's handling of traffic situations, thereby increasing the risk of accidents. Together with Scania, Smart Eye and Viscando, we will get the drivers to trust the system significantly more and use it. They should enjoy using it, be satisfied with it, while also having reasonable expectations of what they can get out of it."
Systems tailored to the driver
The University of Skövde's scientific contribution primarily revolves around AI technology development in relation to human cognition and the driver's behavior concerning the system. Regarding driver behavior, gender is also an important aspect, explains Paul Hemeren.
"Different drivers react differently to arising situations. For instance, clear scientific results show that women and men react differently to various traffic situations. We believe that future safety systems will be more flexible and capable of adapting to who is driving."
Next generation of advanced driver assistance systems
Through the project, researchers will work towards creating the next generation of advanced driver assistance systems. The focus will be on developing methods, called algorithms, to predict and understand various road situations. Simultaneously, researchers will create ways for the car to communicate with the driver in new and smart ways. They will use already collected data from cars on the road and smart infrastructure-based sensors, and study driver behaviors via in-car cameras and driving simulators.
"We will develop significantly better interfaces than what we see today. Information should be quickly displayed for drivers who trust the system and instantly understand what needs to be done or trust that the system will execute an appropriate maneuver. In a broader perspective, we aim to contribute to knowledge that fosters greater trust in intelligent vehicles and intelligent technologies where individuals need support," says Paul Hemeren.
The project is entirely funded with 15 million SEK of which approximately 4.8 million will go to the University of Skövde, which will collaborate with Scania, Smart Eye, and Viscando in the project. The project is ongoing until 2026.
Read more about the project
More about smart cars
Teaches self-driving cars to think like a human
