Designing Feedback to Support Driver-Automation Coordination

点击翻译按钮获取中文摘要

Vehicles are becoming highly automated. Within certain limits, vehicles are now capable of detecting and reacting to hazards as well as maintaining lateral and longitudinal control; they can act in partial, high, or full automation, assuming some or all aspects of vehicle control. Further, drivers are inundated with information. The rapid development of sensor, wireless, and computing technology has given rise to a range of devices (e.g., smart phones), which can entertain and inform drivers, but also distract them. The objective of the proposed research is to support driver-automation coordination in this high information-access environment. Although automation research in other domains (e.g., aviation) can provide some guidance, driving is characteristically different and relevant research in driving is only emerging now.Theme 1 will focus on individual differences. We will investigate how drivers with different personality characteristics, cognitive abilities, and driving experience respond to various automation levels (e.g., distraction engagement, awareness of automation state). To this end, driving simulator experiments coupled with qualitative and quantitative assessments of individual differences will be conducted. Based on our findings, we will develop and evaluate feedback strategies (e.g., feedback about automation state or driver's automation reliance behaviour) that can enhance driver-automation coordination across a range of driver characteristics. Theme 2 will focus on driver attention. In addition to individual differences that are relatively stable over time, different levels of information load experienced by drivers at a given time (distraction) as well as over a period of time (fatigue) can also alter driver-automation coordination. Simulator experiments will investigate how drivers interact with different automation levels under varying information load. Based on our findings, we will develop and evaluate real-time feedback strategies to mitigate the potential negative effects on driver-automation coordination (e.g., situation awareness loss) that can arise from distraction and fatigue. Theme 3 will validate our simulator findings by collecting real world data using our instrumented vehicle, which is one-of-a-kind in Canada. Such validation is rarely undertaken due to significant effort and equipment required, but is very important. Although simulators provide a safe, controlled, and reproducible test environment, they also lack real crash risk. Our results will inform three sets of stakeholders, whom we currently interact with within Canada and internationally: vehicle manufacturers, policy makers, and the public. Just as importantly, this program will train highly qualified personnel with a unique skill set combining expertise in human factors, traffic safety, driving simulators, instrumented vehicles, and statistics.