Motor vehicle crashes represent a significant public health problem in the United States. Per mile traveled, fatal crash rates are elevated for both older and young drivers (Insurance Institute for Highway Safety, 2016). The increased crash and injury risk of older drivers has been attributed to age-related declines in abilities important for driving, as well increased fragility and frailty (Boot, Stothart & Charness, 2014; Dickerson et al., 2007; Meuleners, Harding, Lee & Legge, 2006). The elevated crash risk among younger drivers is generally considered to result from inexperience and immaturity (Romer, Lee, McDonald & Winston, 2014). Efforts to reduce crashes have been multifaceted, focusing on vehicle and road improvements, as well as changes in driver behavior. Of particular note in the vehicle arena are advances being made in automated vehicle technology leading to vehicle designs in which at least some aspects of a safety-critical control function (e.g., steering, throttle, or braking) occur without direct input from the driver (National Highway Traffic Safety Administration, NHTSA, 2016). These advances hold promise for increasing vehicle safety and reducing fatal and non-fatal injuries, particularly among the more vulnerable or high-risk segments of the driving population. However, at least during the early stages of automation adoption, vehicle automation can lead to new and yet unstudied types of risks and errors in drivers or operators. Given the differences in driving abilities, skills, and behaviors of the older and younger novice driving population from other segments of the driving population (Hakamies-Blomqvist et al., 1999; Ponds et al, 1988), these potential errors and risks can have a different impact on safety and thus warrant closer investigation. Important safety issues related to the role of the operator remain and there exists an acknowledged research gap in the understanding of human behavior and interaction with automation (Merat & Lee, 2012), especially in the context of age and experience. An important example of this gap is the fundamental human factors question on the issue of transitioning, or transfer of control, between automated control and manual control in an automated vehicle, especially in Level 2 and Level 3 automated vehicles characterized by some self-driving automation but not the full self-driving automation anticipated for Level 4 (see NHTSA, 2016 for more detail on levels). The overall objective of this exploratory research was to examine specific human factors issues associated with transfer of control to characterize age-related differences in behaviors and reactions to this transition. Behavioral issues and challenges related to automated vehicles are only beginning to be recognized and investigated. The research results from this project provide new insights into age-related differences in the transfer of control between the driver and the automated vehicle with regard to expectations, trust, acceptance, and performance, and how such differences might affect safety. Thus, the knowledge gained from the project will contribute to improving highway safety, an integral component of MTC’s mission. The project had several specific aims including to: 1) characterize driving behavior and responses to transfer of control in an automated vehicle (from automated control to manual control and vice-versa) for older drivers, novice teen drivers, and a comparison group; 2) examine visual scanning behaviors in the three groups during automated driving versus manual driving; 3) assess and compare perceived workload for the three groups when operating an automated vehicle; and 4) explore participants’ perceptions about driver expectations, trust, acceptability, and performance as they relate to automated vehicle transfer of control. The study employed an independent measures design with three groups based on driver age (novice teen drivers age 16-19, older drivers age 65-75, and a comparison group of drivers age 25-45). Participants in each group were presented with simulated driving environments that contained multiple mode transition scenarios (automated mode to manual mode and vice versa). Each participant completed a structured interview immediately following the simulated drive to explore self-perceptions related to the drive. Finally, they completed a self-administered questionnaire to obtain background information. Three groups of drivers were recruited using various techniques (see Appendix A for recruitment materials). The goal was to have enrolled 24 participants in each group at study completion (with attrition estimated at 10% due to simulator sickness and other reasons for withdrawal). The first group was comprised of novice drivers age 16-19 who held a Michigan Level 2 provisional driver license (which allows independent driving with some restrictions) and had their licenses less than 6 months. The second group was comprised of older drivers age 65-75 who had had a regular license for at least 12 months. The third group was comprised of comparison drivers age 25-45 who had had a regular license for at least 12 months. All participants had to drive at least twice a week on average to be eligible for enrollment (based on self-report). Each group was balanced for sex to the extent possible. Teen assent and parental consent were obtained for the participants under age 18 and consent was obtained from those over age 18. Participants were provided with an incentive of $50 for roughly 1.5 hours of study participation (1 hour of simulated driving and .5 hours for the structured interview and background questionnaire). (Author/publisher)
Abstract