Virtual driving with different motion characteristics - braking manoeuvre analysis and validation

Kurzfassung

Driving simulators can be used to examine dangerous situations and new systems without risks to the test driver. Driving situations can moreover be systematically varied and presented in a simulator. However, the validity of the results with regard to real driving must be established. From the different aspects of driving behaviour, this paper focuses on longitudinal control of the car as one of the basic tasks that drivers have to perform. The most common driving strategy in real-life stopping behaviour is the utilisation of a constant deceleration rate. The question we address is: Do drivers behave similarly in simulator experiments? Reymond et al demonstrated the important role of kinaesthetic cues. They proved that motion response generally increases the validity of driving behaviour in virtual environments. Siegler et al conducted a small braking experiment and found that braking strategies remained stable, and that decelerations were within realistic ranges in the presence of motion cues. Further, the drivers were able to stop at a specific given point of the road. A more comprehensive approach was used by Boer et al with the scope of analysing time sequences instead of statistical characteristic values of a complete braking or stopping manoeuvre. A set of useful decision- and control-based variables were introduced and the results were validated with real-world driving data. In later analysis, they tried to find the optimal scaling factor and its general valid function based on the biases in distance, speed and acceleration perception. The insights obtained by these former studies were used to intensively analyse the data of an experiment conducted at the DLR Institute of Transportation Systems. In order to evaluate the participant’s ability to control a car within the new dynamic driving simulator (6 DoF), the driving behaviour in the simulator is compared to driving in the real world with an equipped car (the DLR ViewCar). The influence of different motion platform characteristics on the validity of the braking behaviour is investigated and presented in this paper.