Contents lists available at ScienceDirect Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap Dangerous intersections? A review of studies of fatigue and distraction in the automated vehicle Gerald Matthews a, ⁎ , Catherine Neubauer b , Dyani J. Saxby c , Ryan W. Wohleber d , Jinchao Lin d a Institute for Simulation and Training, University of Central Florida, 3100 Technology Pkwy, Orlando, FL, 32826, United States b University of Southern California, United States c Medical College of Wisconsin, United States d University of Central Florida, United States ARTICLE INFO Keywords: Safety Driver behavior Fatigue Stress Automation Distraction ABSTRACT The impacts of fatigue on the vehicle driver may change with technological advancements including automation and the increasing prevalence of potentially distracting in-car systems. This article reviews the authors’ simu- lation studies of how fatigue, automation, and distraction may intersect as threats to safety. Distinguishing between states of active and passive fatigue supports understanding of fatigue and the development of coun- termeasures. Active fatigue is a stress-like state driven by overload of cognitive capabilities. Passive fatigue is produced by underload and monotony, and is associated with loss of task engagement and alertness. Our studies show that automated driving reliably elicits subjective symptoms of passive fatigue and also loss of alertness that persists following manual takeover. Passive fatigue also impairs attention and automation use in operators of Remotely Piloted Vehicles (RPVs). Use of in-vehicle media has been proposed as a countermeasure to fatigue, but such media may also be distracting. Studies tested whether various forms of phone-based media interacted with automation-induced fatigue, but effects were complex and dependent on task configuration. Selection of fatigue countermeasures should be guided by an understanding of the form of fatigue confronting the operator. System design, regulation of level of automation, managing distraction, and selection of fatigue-resilient personnel are all possible interventions for passive fatigue, but careful evaluation of interventions is necessary prior to de- ployment. 1. Introduction Threats to safety often arise when multiple vulnerabilities converge (Reason, 2000). Fatigue, distraction, and system automation, may all elevate risk to vehicle operators, at least in certain contexts. Williamson et al. (2011) identified three safety-relevant factors associated with fatigue: sleep homeostasis factors, circadian influences and task effects. Their review concluded that sleep loss has the strongest effects on crash risk, but more research is needed on task-induced fatigue. The dangers of distraction have become more salient with increasing use by drivers of in-car systems, including informational systems such as GPS, en- tertainment systems, and cell phones, which allow the driver to engage in extraneous activities (Strayer, 2015). Fatigue and distraction factors have been extensively studied in conventional vehicles, but there is only limited research on how they may interact in their effects on safety. Both full and partial vehicle automation can bring a range of safety benefits. Automation refers to multiple systems (Banks et al., 2014). Allocation of basic tasks such as control of speed to automation may mitigate driver workload, whereas collision warning or avoidance systems can prevent specific driver errors. Safety concerns about system automation derive from the loss of situation awareness (SA) that may occur in some circumstances (Strand et al., 2014; Young and Stanton, 2007). SA refers here to the driver’s internal model of the traffic en- vironment and its safety implications. In addition, future systems may increasingly require active collaboration between the driver and the automated systems, mediated by driver initiation of automated func- tions (Banks and Stanton, 2016). Optimization of trust in the automa- tion is critical: under- or over-dependence on the automation may im- pair performance (Lee and See, 2004; Parasuraman and Riley, 1997). Research is lacking on whether the safety impacts of fatigue and distraction in the automated vehicle are similar to those observed in conventional vehicles. We cannot necessarily summate risk factors to predict that, crudely, risk = fatigue + distraction + automation-in- duced SA loss. For example, distraction in the form of conversation may actually help to maintain alertness in the fatigued driver (Atchley and https://doi.org/10.1016/j.aap.2018.04.004 Received 21 July 2017; Received in revised form 4 April 2018; Accepted 5 April 2018 ⁎ Corresponding author. E-mail address: gmatthews@ist.ucf.edu (G. Matthews). Accident Analysis and Prevention xxx (xxxx) xxx–xxx 0001-4575/ © 2018 Published by Elsevier Ltd. Please cite this article as: Matthews, G., Accident Analysis and Prevention (2018), https://doi.org/10.1016/j.aap.2018.04.004