The DRAMA project develops knowledge within driver activity mapping to improve interaction between semi-automated and automated vehicles and human drivers and passengers - with the overall goal to improve traffic safety and driver/passenger comfort. With correct classifications of what the driver is doing, the human-machine interaction can be directed to the most suitable modality (visual/audio/haptic) at each moment. If the car knows the full body position (sitting, lying, twisted to the back-seat, leaning to the side, etc) of its passengers, the safety functions can be adapted to the in-the-moment best deployment of for example airbags, steering, brake and crash avoidance patterns. The decision if the vehicle can do a hand- over, or should perform a safe stop can be determined based on in-vehicle activity and driver attention/disengagement in the driving task.
We expect to find methods/algorithms/tools that can be used to map driver body posture/facial expressions etc. and to combine them in models that can describe the driver behaviour/activity. The project will also perform research and implement a concept for a camera-based data acquisition system to be able to capture the behaviour.