RE-ENGAGE – Human activity detection to re-engage drivers in autonomous cars

One of the advantages with  highly automated vehicles (SAE level 4) is that the driver can disconnect from the driving duties and studies have shown that people do engage or would like to engage in various activities, such as socializing, work or just sit back and relax and take a powernap in autonomous mode. The car’s user interface information and driving actions are important factors when it comes to how the driver experience and trust the automation during different activities. Adapting the car’s user interface information to the activity at hand has the potential of making the ride safe and more pleasurable. In addition, it gives a possibility to re-engage the driver when immersed in non-driving related activities. 

This will be done by using machine learning algorithms, able to recognize different activities from video data (i.e. reading a book and taking a powernap), in turn influencing the interface information provided to the driver. The non-driving related activities in focus in this project are reading a book and taking a powernap, since they are similar from a detection perspective, but they require very different feedback to create a correct driver action. From UX perspective we will use a novel approach called Soma design to create prototypes that activates the whole body through multimodal interactions.

RISE is the main applicant and coordinator of the project with Volvo Cars and Smart Eye AB as the industrial partners. The project collaborates with Stanford School of Medicine & Human Factors, Center for Interdisciplinary Brain Sciences Research, but they do not have any economical part in this project and financing of their research work is applied from the strategic collaboration with Volvo Cars.