The mission of the project is to provide for the development of artificial intelligence systems that understand and manage the state of the driver to create a safe and enjoyable experience in semi-autonomous vehicles. At the core of the research effort is the development and real world evaluation of deep learning algorithms that enable effective communication and transfer of control between human and machine. This includes sensing driver gaze, emotion, cognitive load, drowsiness, hand position, posture, and fusing this information with the perception of the driving environment to create safe, reliable vehicles that drivers can learn to trust.
“We are looking to lead the world in the application of state-of-the art deep learning methods for semi-autonomous vehicle systems in a way that ensures the human being is the core focus of every algorithm and interface we build," says Lex Fridman, MIT. "It is clear that the global focus on autonomous vehicles must begin to consider increased investment in human centered vehicle system that support appropriate driver engagement through trusted safety and an enjoyable mobility experience," says Bryan Reimer, MIT. “We believe that drivers will be traversing the world’s roadways for decades to come as higher levels of automation steadily increases.”
“Still, every year 1.4 million people die in traffic fatalities. The introduction of driver assistance technology is not enough. To save more lives, we need to establish trust between the driver and the car’s intelligence. I am confident that this collaboration with leading researchers will accelerate the industry’s ability to deliver future safety solutions,” says Ola Boström, VP Research at Autoliv.
Learnings from the research collaboration will likely be used in the development of software from Zenuity, the software joint venture of Autoliv and Volvo Cars.