Advanced Driver Vehicle Interface in a Complex Environment
Noldus, together with partners HAN, TomTom, TNO, Delft University and Green Dino, takes up the challenge to build a driver observation system for human-machine interface (HMI) evaluation studies.
The concern for the impact of our mobility systems on safety, environment and congestion has resulted in many innovations in vehicles and infrastructure in the last decades. It is absolutely clear that the driver plays a crucial role in this. Putting the gas pedal to the ground, not allowing sufficient distance, and taking risks on the road by aggressive driving behavior are all factors being controlled by the driver. For that reason, engineers have tried to find innovative solutions by giving better support to the driver. Regarding safety, over 80% of all accidents are still caused by human error, being primarily a result of poor recognition and inappropriate decision making. It is therefore remarkable that the type of driver and the driver state are hardly taken into account in these systems.
We can distinguish between young drivers (responsible for 20-30 % of all driver deaths), elderly drivers, truck drivers, etc. The driver state refers to mental workload, fatigue, alertness, drowsiness, driving skills, and so on. It seems obvious that better (that is, more effective) driver support may be obtained by adding driver state information to vehicle state data and information about traffic and road conditions, as input to the support system. This is referred to as DSE (Driver State Estimation) where we focus on workload estimation. However, DSE is alone is not sufficient. Being able to prove its effect is equally important. The driver observation system for human-machine interface (HMI) evaluation studies that Noldus and its partners are building is the first generation of a standardized validation tool. Its design is highly adaptable in order to support various automotive research questions in instrumented vehicle tests and naturalistic driving studies. The system can collect and interpret information from various sources, such as actual or simulated vehicles, portable navigation devices, cameras, eye trackers, and psycho-physiology.
This project has resulted in the product DriveLab, which is now available from Noldus.