N. Teasdale, V. Cantin, G. Desroches, J. Blouin, M. Simoneau
Proportionally more accidents are occurring in complex situations, such as intersections. Presumably, this higher rate of accidents is associated with a higher momentary mental workload created by the intersection. Two experiments are presented to examine if this mental workload can be manipulated while driving in a simulator. We adopted the probe reaction time technique to measure the attentional demands while driving. For each experiment, ten active drivers drove through a continuous 10-km scenario including rural and urban sections. They were asked to follow speed limits and to comply with local traffic laws throughout the course of the experiment. They were also told to consider driving as the primary task. For the first experiment, auditory stimuli were given while driving on open roads and when approaching intersections. The stimuli could be presented when the car was 100 m or 60 m from the intersection. For the second experiment, the stimuli were given while driving on open roads, when approaching intersections (75 m), and when performing lane change and doubling maneuvers. RT (ms accuracy) was defined as the temporal interval between the auditory stimulus and the onset of a verbal response (a loud top, detected from the analog signal of a piezo-electric microphone fixed on a headset). RT was also measured for a baseline control condition without driving (10 stimuli). Several variables characterizing the driving performance were collected. For both experiments. driving yielded RTs that were slower than for the baseline condition. More important, experiment 1 showed that stopping at an intersection required more cognitive demands than driving on open roads (on average, 519 ms for the open road conditions vs 727 ms for both intersection conditions); experiment 2 yielded similar results and the lane change and doubling maneuvers was the driving context that required the most important cognitive demands (on average, 612 ms for the open road conditions, 675 ms for the intersection conditions, and 816 ms for the lane change and doubling maneuvers). Results clearly show that driving, even for ideal conditions, is not an automatic task. Future experiments should attempt to examine whether various categories of drivers (eg novice, elderly) are affected differentially by various contexts and whether the increased mental workload has an effect on the driving performance. The technical approach developed could allow to develop more challenging scenarios and to validate workload observed for real driving contexts.
Keywords: simulator; attention; intersections; mental workload