The effects on driver behaviour, workload, and acceptance, of different levels of task allocation between a collision avoidance system (CAS) and the driver, were studied in an advanced driving simulator. Driving performance was described in terms of time headway, speed level, lateral position, and left lane driving (including overtaking behaviour). Workload was measured with the NASA-TLX questionnaire, acceptance with attitude questions. Forty subjects, between 23 and 58 years and experienced as drivers, were randomly assigned to four experimental conditions. It was found that a CAS system capable of taking over the control of the car produced the most favourable effects on drivers headway, and also reduced the number of overtakings. But, seen from the drivers point of view, that system was regarded as the most intrusive, and also most disturbing system. It was recommended that a GIDS, CAS system should either give a warning, or suggest an appropriate action. It was also found that the use of the intelligent gas pedal, as a link between driver and CAS, did not lead to an increase of the drivers' workload. Finally, some consequences of the time to collision criteria were discussed. This is a reprint from DRIVE project V1041 Generic Intelligent Driver Support Systems (GIDS), Sept 1991. (A)
Samenvatting