The goal of this conference was to provide opportunities for participants to exchange novel ideas and professional experiences, and discuss case studies and research applications. Presentations featured a variety of research areas and activities including vehicle automation and connectivity, shared mobility, enabling technology, human factors, policy and planning, and infrastructure design and management.
Huang presented his work titled, "Predicting Mind Wandering during Semi-autonomous Driving and Exploring Potential Mitigation Strategies," and won one of two Outstanding Speaker Awards for Best Presentation. Liang also presented his research titled, "The Effect of Secondary Cognitive Task Difficulty on Headway Maintenance and Perceived Workload with Lane Keeping Systems."
NHanCE Lab member and undergraduate SURF student Alec Gonzales also attended the conference with Huang and Liang, held at Purdue's Discovery Park on May 31, 2019.
ABSTRACT - "Predicting Mind Wandering During Semi-Autonomous Driving and Exploring Potential Mitigation Strategies"
G. Huang, B.J. Pitts
Significant growth in the number of autonomous vehicles is projected in the coming years. Expected benefits of autonomous vehicles include increased mobility for the general population and improved public roadway safety. With the increasing reliability of autonomous vehicles, drivers are more likely to become distracted and/or disengaged from the driving task. One major concern regarding the attention of drivers is internal distraction, most notably, mind wandering, i.e., a shift of one’s attention from the external world to internal thought processes. Findings from the literature in this area suggest that mind wandering can lead to a reduction of visual range and environment (and situation) awareness, longer response times to driving-related events, shorter headway distances, and higher driving velocities in the manual driving context.
However, little work has been done to understand its impact on autonomous driving, which is important given that current automated vehicle functions can fail as a result of complex construction areas, high traffic volume, and/or missing lane markers. In these cases, drivers may need to be ready to take over control from the vehicle at any time. In addition, physiological measurements, such as pupil size, heart rate, and/or skin conductance level, may be used to predict mental states. Yet, it is unclear which particular techniques can differentiate between mind wandering and attentive states. Monitoring mental states and driving performance are both prerequisites to designing effective warning signals that take drivers out of a mind wandering state and bring them back into the loop. Multisensory interfaces may serve as a feasible approach to alerting systems, however, more research is needed to determine their potential benefits. Given these research gaps, this work aims to take initial steps towards (a) quantifying the effects that mind wandering semi-autonomous driving performance, (b) developing a model to predict when a person is mind wandering based on driving performance and physiological data, and (c) investigating the effectiveness uni-, bi-, and trimodal combinations of visual, auditory, and tactile cues to represent takeover alerts.
ABSTRACT - "The Effect of Secondary Cognitive Task Difficulty on Headway Maintenance And Perceived Workload With Lane Keeping Systems"
N. Liang, B.J. Pitts
Vehicle automation is developing at a rapid rate worldwide. However, even lower levels of automation, such as SAE Level-1, are expected to reduce drivers’ workload by controlling either speed or lane position. At the same time, however, drivers’ engagement in secondary tasks may make up for this difference in workload displaced by automation. Previous research has investigated the effects of adaptive cruise control on driving performance and workload, but little attention has been devoted to lane keeping systems (LKS). In addition, the influence of secondary cognitive tasks on Level-1 driving performance is also not well understood. The goal of this study was to examine the effects of secondary cognitive tasks on driving performance and perceived workload while using LKS. Nine participants drove a simulated vehicle in manual and LKS modes, while maintaining a specific headway and performing a secondary cognitive (n-back) task with varying levels of difficulty. Results showed that standard deviation of headway and NASA-TLX workload scores were significantly higher during the most difficulty secondary task. Also, LKS was not found to improve driving performance nor reduce perceived workload. This paper highlights potential performance costs and benefits associated with LKS technology and proposes directions for future research.