Implementation of Driving Simulator with Eye Tracking Device for Stroke Patient in Resumption of Driving

In order to establish an environment for behavioral evaluation during car driving, we developed a novel drive simulator that consists of a head/eye tracking device and custom-made 3D CG modeling based on the drone-captured test-driving field. With the use of head tracking-based view manipulation, users can get a wider field of view even though the system only has a single display. The developed drive simulator has been tested for evaluation in total 60 stroke patients including visuospatial neglect, attention deficit, aphasia, and hemianopia. A 300m straight road was placed in the virtual space, and various elements were involved in the following tasks: (1) simple straight, (2) signal task, (3) number-call task, and (4) pedestrian detection task. Most of patients showed symptom-dependent errors, such as, excessive speed, failed detection of signal, sign, and pedestrian. The gaze and head data clearly showed a bias due to the neglect and visual deficit in typical two patients. Since head/eye tracking data and recordings of the steering and pedals give us plenty of information for car driving, our developed system would effectively work for the detection of whether patients have the potential to re-issue driving licenses or not.