Tesla Autopilot Through Constructions: Investigating the Effect of On-Road Partially-Automated Driving through Construction Zones

Tesla Autopilot Through Constructions: Investigating the Effect of On-Road Partially-Automated Driving through Construction Zones | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Tesla Autopilot Through Constructions: Investigating the Effect of On-Road Partially-Automated Driving through Construction Zones Francesco Biondi, Praneet Sahoo, Noor Jajo This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4675940/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 12 Mar, 2025 Read the published version in Scientific Reports → Version 1 posted 14 You are reading this latest preprint version Abstract Partially-automated driving systems are designed to control the vehicle’s speed and acceleration without input from the human driver on the condition that the driver maintains alertness. These systems are promised to make driving safer especially when driving in road sections exhibiting a higher risk of collisions like construction zones. Despite this, little knowledge is available on how these systems are used in these accident-prone areas and the effect they may have on drivers’ workload and glance allocation. This study aims to fill this gap by having participants drive a Tesla vehicle in Autopilot and manual mode through three road sections: pre-construction, construction, and post-construction. Results show no differences in cognitive workload by driving mode or construction zone. An increase in glances directed away from the forward roadway toward the vehicle’s touchscreen was observed during partially-automated driving in the pre-construction zone, a pattern that, notably, continued on when driving throughout the construction zone. These findings adds to the literature on the human factors of partial automation. More importantly, because drivers failed to increase the amount of time looking at the forward roadway when entering the construction zone, they show the perniciousness of partially-automated driving and the detrimental effect these systems may have on safety. Biological sciences/Psychology/Human behaviour Earth and environmental sciences/Environmental social sciences/Psychology and behaviour Tesla Autopilot Cognitive Workload Glances Road Safety Partial Automation Figures Figure 1 Figure 2 Introduction Partially automated driving systems, also known as SAE level-2 automation, are systems capable of maintaining the lateral and longitudinal control of the vehicle in certain road and traffic scenarios (SAE, 2021 ). Unlike lower or higher levels of automation wherein it is clear who is in control of the vehicle – the human driver in SAE level-0 and 1, and the automated system in SAE level-4 and 5 – in partial automation responsibilities are shared between the two. In particular, when the automated system is in control of the vehicle’s steering and accelerating, the human driver is still responsible for monitoring the system functioning and resuming control whenever necessary. The introduction of this technology has been promised to have a positive effect on safety especially in situations where the driver’s workload may be higher due to conditions like congested traffic or poorer visibility. Recent estimates point the adoption of this technology will be able to prevent 1 every 4 crashes and fatal injuries, and 1 every 3 fatalities in the year 2050 (Naumann et al., 2023 ). Despite such rosy forecasts, the shared vehicle control between the human driver and the automated system is thought to be at the center of numerous fatal accidents. Recent investigations by US regulators point at driver complacency and misuse as contributing factors to these accidents. For example, the National Transportation Safety Board (NTSB) investigation on the 2016 fatal accident, wherein a partially-automated vehicle crashed into a tractor semitrailer in Florida, reports that the absence of design constraints that limited the use of these systems only when appropriate could give rise to drivers misusing these vehicles (NTSB, 2017 ). In a report following a similar crash in 2018, NTSB also points at the vehicle’s ineffective monitoring of driver engagement, which resulted in driver inattentiveness and complacency, as a probable cause of the accident (NTSB, 2018 ). The 2023 NHTSA and Transport Canada recall on Tesla’s Autosteer system also indicate that, because of its design, drivers may misuse the partially automated system which could lead to an increased risk of collisions (NHTSA, 2023 ) The conclusions reached in these investigations (also see NTSB, 2020 ) find support in the broader literature on partial automation use. Studies largely point at the higher risk of driver disengagement and distraction occurring whenever the automated system is engaged. Prior work by Biondi et al. (Biondi et al., 2018 ; Biondi & Jajo, 2024 ) investigated differences in drivers’ workload and attention allocation when transitioning from manual to partially-automated driving. While mental workload seemed largely unaffected by driving mode, drivers became less attentive toward the driving scene as the control of the vehicle was handed over to the automated system. When compared to manual driving, such attention decline appeared to peak over time during partial automation, an effect that seemed consistent across distinct level-2 systems (Biondi et al., 2023 ). Findings by Morando et al. ( 2021 ) and Yang et al. ( 2021 ) with drivers operating a Tesla vehicle in Autopilot mode point at similar patterns. In the time preceding the transition to manual driving, Morando et al. observed a higher frequency of longer glances directed toward the vehicle’s center-stack. The study by Yang et al. also highlights that lesser engagement in the driving task when the Autopilot system is on degraded the driver’s ability to resume manual control of the vehicle post-transition. Naturalistic data by Noble et al. ( 2021 ) align with this literature showing a higher prevalence of cellphone use during partially-automated driving. These findings warn caution about the use of these systems, especially in conditions where the risks of vehicle-to-pedestrian or vehicle-to-vehicle collisions are heightened. Construction zones represent particularly perilous road sections. In 2022, 891 people were killed and 37,701 were injured in construction zone crashes in the US alone (NSC, 2024 ). Work by Khattak et al. ( 2002 ) shows that the risk of crashes during construction zones is 21.5% higher than that in pre-construction zones, with the formers also exhibiting a 23.8% and 17.3% increase in non-injury and injury crash risks, respectively. A more recent analysis conducted by Lym and Chen ( 2021 ) indicate that the severity of vehicle crashes resulting from driver distraction tends to be higher in construction zones. Consistent findings are offered by Roshandeh et al. ( 2016 ) who also observed an increased risk of hit-and-run crashes in construction zones in the presence of a distracted driver. With construction zones being so dangerous, the adoption of partially automated systems holds the potential of benefiting safety in these collision-prone road sections. Yet little evidence is available on the effect that using these systems has on driver behavior when driving in construction zones. Related literature has in fact largely focused on investigating driver interaction with these systems in favorable traffic and road conditions (Banks et al., 2018 ; Gaspar & Carney, 2019 ; Yang et al., 2021 ). With this in mind, the current study aims to fill this gap by having participants drive a Tesla vehicle in either partially automated or manual mode in three distinct zones: pre-construction, construction, and post-construction. The first objective is to investigate the combined effect that driving mode and zone has on cognitive workload. Prior research has shown mixed findings about the effect of level-2 automated driving on cognitive workload with studies either showing no differences between partial automation and manual driving (Lohani et al., 2021 ; Mcdonnell et al., 2023 ) or a decrease in cognitive workload in level-2 mode (Biondi et al., 2018 ; Solís-Marcos et al., 2017 ). With the driving task arguably becoming more demanding during construction zones (Ma et al., 2023 ; Shakouri et al., 2018 ), we expect cognitive workload to increase when transitioning from the pre-construction zone into the construction zone in both manual and level-2 mode. The second objective is to investigate the combined effect of driving mode and zone on glance allocation. Prior research has shown glance allocation toward driving-unrelated areas to increase during level-2 driving (Morando et al., 2021 ; Yang et al., 2021 ). In the current study we also expect a similar pattern during level-2 driving, but with glances toward the forward roadway possibly increasing with the beginning of the construction zone. Method Participants Twenty-one volunteers were recruited from the University of Windsor. Their average age and standard deviation of age were 22 years and 4.36 years, respectively. All participants were fluent English speakers, had normal or corrected-to-normal vision and hearing, held a valid driver’s license, had proof of car insurance, and had not been the at-fault driver in an accident within the past 2 years. Participants were also required to complete 30-minute defensive driving course and be affiliated with the University of Windsor. A University of Windsor Research Ethics Board approval (REB #20–141) was obtained for the study. All methods were performed in accordance with the relevant guidelines and regulations. Experimental Design A factorial design with two independent variables were used in this study. The within-subject independent variables were driving mode (2 levels) and zones (3 levels). Participants drove a vehicle in either manual or level-2 automation mode, and through three zones: pre-construction, construction, and post-construction. Dependent measures included the performance to the ISO Detection Response Task (ISO, 2015 ) and the percentage of time spent looking at each of the following areas of interest (AOI): forward roadway, vehicle’s touchscreen, side mirrors, and rearview mirror. Equipment and Data Processing Vehicle A 2022 Tesla Model 3 was used for the study. Participants drove the vehicle in manual and level-2 mode. In level-2 mode, the vehicle was driven with both Adaptive Cruise Control (ACC) and Lane Keeping Assist System (LKAS) engaged. ACC maintained the vehicle at a set speed and distance from the vehicle in front and, during the study, it was set to the maximum distance of 7 car lengths. LKAS maintained the vehicle within the lane and, during the study, drivers were instructed to occupy the right lane when two lanes were available or the middle lane when three lanes were available. This was done to minimize lane changing and the need to overtake slower traveling vehicles. When working together, ACC and LKAS meet all requirements necessary of SAE level-2 automation.In manual mode, participants drove the vehicle manually with no assistance from either ACC or LKAS. Route Two routes were considered for the study. A familiarization route was used for participants to familiarize with the vehicle and the level-2 system. Participants drove a loop through Huron Church road and Ontario Highway 3 between approximately the University of Windsor campus and the St. Clair College campus in Windsor, ON. The experimental route consisted of a section of Ontario Highway 401 between Exit 13 in Windsor, ON and Exit 81 near Chatham, ON. The road consists of a combination of three-lane and two-lane dual carriageways that provides sufficient space between vehicles. The daily traffic volume on the route was approximately 25,000 vehicles. For the current study, three distinct zones were identified: pre-construction, construction, and post-construction (Fig. 1 A). The pre-construction zone consisted of the road section from the start of the drive (kilometer 13) to the beginning of the construction zone (kilometer 50). The start and end of the construction zone were marked by a road sign (Fig. 1 B) and the presence of the last orange cone (kilometer 65; Fig. 1 C), respectively. The post-construction zone followed until the end of the drive at kilometer 81. Detection Response Task The vibrotactile version of the Detection Response Task (DRT) manufactured by Red Scientific Ltd (Salt Lake City, UT, USA) was used in the study. A vibrotactile motor was placed on the inside of the participants’ left elbow area and a microswitch was attached to either the index or middle finger of the left hand. A vibration, similar to that emitted by a cellphone, was presented every 3 to 5 seconds with a duration of 1 second. Participants were instructed to respond to the vibration by pressing a microswitch as fast as possible. Reaction times (RT) in milliseconds and hit rates were recorded. RT were recorded as the time interval between the onset of the vibrotactile stimulus and the depression of the microswitch. RT faster than 100 ms or longer than 2,500 ms were eliminated from the calculation. Nonresponses or responses produced later than 2,500 ms were considered as misses. Average RTs were calculated by driving mode and zone. Cameras The vehicle was equipped with three GoPro HERO8 Black cameras. The driver view camera was points at the driver and footage from this camera was used to record eye glances and eye movements. The front view camera was pointed at the forward roadway. The touchscreen view camera was pointed at the vehicle’s touchscreen. The cameras recorded at 1080p at 240 frames per second. Glances Footage from the driver camera was manually coded by two trained coders. Four areas of interest (AOI) were identified: forward roadway, vehicle’s touchscreen, side mirrors, and rearview mirror. Glances directed toward the forward roadway included any glance directed toward the center, the left, or the right side of the road to, e.g., inspect the road or vehicle in front, or check for potential hazards or signs. Glances directed at the vehicle’s touchscreen included any glance directed at the touchscreen displaying information about the vehicle’s functioning. Glances directed at the side mirrors and rearview mirrors included any glance directed toward these areas to, e.g., check for the surrounding vehicles. The selection of these AOI is consistent with prior work by Biondi et al. ( 2015 ) and Gaspar and Carney ( 2019 ). Two independent coders were assigned to each video and manually coded it frame-by-frame to measure the time each driver spent looking at each AOI during the manual and L2 drive. Possible discrepancies between the two coders were flagged and reviewed by a third coder. Cohen’s kappa was calculated as a measure of inter-coder reliability. A Cohen’s kappa of 0.85 was found indicating high consistency between the two coders (McHugh, 2012). The percentage of time spent looking at each AOI was calculated as the ratio between the time spent looking at each AOI divided by the total time spent in each of the three zones (pre-construction, construction, post-construction). For example, if a driver spent 30 seconds looking at the forward roadway while driving in a construction zone that took a total of 60 seconds to travel, the percentage of time spent looking at that AOI in that zone would be 50%. This metric was calculated as such to account for differences in zone lengths and the time it took participants to travel through them. Average percentages of time looking at each AOI were calculated by driving mode and zone. Procedure Study intake . Prior to the day of the study, participants were provided with a video tutorial on how to use the level-2 system on the vehicle. This was done to ensure that all participants were provided consistent information about how to use the system. On the day of the study, participants met with the research associate and were directed to where the vehicle was parked. When inside the vehicle, participants were asked to review the signed informed consent form and REB checklist and ask any questions they may have. The research associate then advised them to sit in the car to learn how to adjust the mirrors, seats, and steering wheel. Familiarization phase . During this phase, participants were given 1 minute to practice completing the DRT when the vehicle was stationary. They were also reminded to put their phones on mute, and that the could not make use of their phones for the entire duration of the drive, unless in case of an emergency. Participants drove the familiarization route in manual and L2 mode for as long as they needed, which took up to 20 min. Experimental phase . Once participants verbally confirmed they were comfortable driving the vehicle in manual and L2 mode, the experimental phase begun. Participants drove the experimental route twice: once in level-2 mode and once in manual mode. The order of the two drives was counterbalanced across participants. Each drive took approximately 40 minutes to complete. Participants exited the highway at kilometer 81 near Chatham, ON and parked at a gas station where they could take a 15-minute break. After the break, they reentered the highway in the opposite direction and the second drive begun. The second drive ended at exit 13 in Windsor, ON at which point participants were instructed to drive back to the University of Windsor campus. The experimental phase took up to 2 hours. Data analysis Bayes factor analyses were adopted for data analyses. Unlike the traditional null-hypothesis statistical testing (NHST) which relies on the p-value to determine whether the null hypothesis is accepted, Bayesian analyses set up two competing models, one for the null hypothesis and one for the alternative hypothesis, and estimate which of the two models is more likely to generate the data at hand. Bayesian analyses transform p-values into direct evidence against the null hypotheses (Held and Ott, 2018). A Bayes Factor (BF) is calculated as the ratio between the marginal likelihood of the null model and that of the alternative model (Quintana & Williams, 2018 ). A BF equal to X indicates that the data is X times more likely under the alternative hypotheses than under the null hypothesis. BF range between 0 and infinity. BF less than 0.33 indicate evidence in support of the null hypothesis, with smaller values suggesting stronger evidence for H 0 . BF greater than 3 indicate evidence in support of the alternative hypothesis, with larger values suggesting stronger evidence for H 1 (Dienes, 2014 ). Unlike NHST, Bayesian analyses provide evidence in support of either the null or the alternative hypothesis, as well as offer information about the strength of the evidence. Data processing and analyses were conducted using R (version 4.1.0) and RStudio (version 2023.03.0 (Racine, 2012 )). The tidyverse (version 2.0) and BayesFactor (version 9.12) libraries were adopted for data processing and Bayesian analyses, respectively. Results Results are presented by objectives. Investigate the combined effect that driving mode and zone has on cognitive workload A model was set up with driving mode (2 levels: manual, level-2) and zone (3 levels: pre-construction, construction, post-construction) as the independent factors and DRT RT as the dependent measure. A BF of 0.19 was found for driving mode indicating strong evidence in favor of the null hypothesis. This finding aligns with prior findings by McDonnell et al. ( 2023 , 2021 ) who also found no significant changes in drivers’ cognitive workload between driving in manual mode and driving with the level-2 system engaged. Strong evidence in support of the null hypothesis was also found for the main effect of zone (BF = 0.18) indicating that, although DRT RT increased from the pre-construction to the construction zone, this difference was nonsignificant. A BF of 0.005 was found for the zone by mode interaction indicating no significant interaction between these two factors. DRT RTs are presented in Table 1 . Table 1 Mean and standard error (SE) of DRT RT in milliseconds by zone and driving mode. Zone Pre-construction Construction Post-construction Mean SE Mean SE Mean SE Manual 526 41.9 622 71.3 602 46.8 Level-2 549 54.6 594 52.8 603 48.8 Investigate the combined effect of driving mode and zone on glance allocation Models were set up with driving mode (2 levels: manual, level-2), zone (3 levels: pre-construction, construction, post-construction) and AOI (4 levels: forward roadway, side mirrors, rearview mirror, touchscreen) as the independent factors and the percentage of time spent looking at AOIs as the dependent measure. A BF of 2.95 x 10 642 was found for AOIs indicating differences in the time spent looking at each AOI. Based on this result, separate models were set up, each investigating separate AOIs. For forward roadway glances, a BF of 5.4 x 10 3 was found for mode indicating a strong difference between the two modes. Whereas drivers spent an average of 96.2% (SE = 0.83%) of the time looking at the forward roadway during manual mode, this decreased to 91% (SE = 0.44%) during level-2 driving. A BF of 0.08 was found for zone suggesting that the time spent looking at the forward roadway did not change across zones (pre-construction: M = 94.3%, SE = 0.64%; construction: M = 93.9%, SE = 1.01%, post-construction = 93.5%, SE = 0.98%). Similar analyses were conducted for touchscreen. A BF of 4.2 x 10 3 was found for mode indicating that drivers spent more time looking at the vehicle’s touchscreen during level-2 mode (M = 6.14%, SE = 0.61%) relative to manual mode (M = 2.79%, SE = 0.35%). A BF = 0.07 was found for zone suggesting that the time spent looking at the touchscreen did not change across zones (pre-construction: M = 4.36%, SE = 0.58%; construction: M = 4.26%, SE = 0.68%, post-construction = 4.73%, SE = 0.74%). Analyses conducted for glances directed at the side mirrors revealed a BF of 1.81 for the main effect of mode and a BF of 0.11 for zone suggesting that the percentage of time looking at this AOI did not change across modes nor zones. Similar results were found for the rearview mirror. A BF of 0.64 was found for mode and a BF of 0.08 was found for zone, suggesting no differences in the time spent looking at this AOI by mode or zone. Data for forward roadway and touchscreen AOIs are presented in Fig. 2 . Discussion In this study we set out to investigate two distinct objectives. The first objective was to investigate the effect that driving mode and zone had on drivers’ cognitive workload. Prior research on the topic has produced mixed findings. Studies by McDonnell et al. ( 2021 , 2023 ) and Lohani et al. ( 2021 ) showed no differences in drivers’ workload between the two modes. Conflicting patterns were observed by Biondi et al. ( 2018 ) and Solis-Marco et al. (2017) who found lower mental workload when driving in level-2 mode. Our results align with the work by McDonnell et al. and Lohani et al. with no differences being found in DRT RT between manual and partial automation. No differences in DRT RT were found between the three zones under consideration indicating that cognitive workload did not change across the pre-construction, construction, and post-construction zones. This result is at odds with prior research showing an increase in driving demand when travelling through construction zones (Ma et al., 2023 ; Shakouri et al., 2018 ). Although the ISO DRT has proven sensitive in discriminating changes in cognitive workload induced by diverse driving and road conditions (Biondi et al., 2024 ; Boehm et al., 2021 ; Castro et al., 2019 ), it is still possible that the changes in driving demand elicited by our three zones may have not been sufficient to produce significant changes in DRT performance. Our second objective was to investigate the effect of driving mode and zone on glance allocation. Analyses conducted on glances directed at the forward roadway evidenced a decline in the percentage of time looking at this AOI during level-2 driving. This datum aligns with prior findings by Morando et al. ( 2021 ) and Yang et al. ( 2021 ) evidencing a decline in visual attention toward the forward roadway when the partially automated system is engaged. As the time spent looking at the forward roadway decreased, the time spent looking at the vehicle’s touchscreen increased in level-2 mode. Again, this pattern is consistent with prior work by Noble et al. ( 2021 ) who also suggest that driving in partially automated mode increases the likelihood of engaging in driving-unrelated activities. Analyses investigating changes in glance allocation by zone revealed important patterns. No differences in the time spent looking at the forward roadway was found between pre-construction, construction, and post-construction zones, indicating that the presence of the construction zone didn’t bring drivers to increase the time spent looking at the forward roadway. This finding highlights a dangerous pattern. It was indeed expected that as drivers transitioned into the construction zone, this may have increased the amount of forward glances executed to anticipate the presence of hazards and construction workers. The fact that, during the construction zone, drivers kept glancing away from the road toward the vehicle’s touchscreen to the same extend they did in the pre-construction zone suggests that the potential for distraction resulting from level-2 system (see Noble et al., 2021 ) might not be limited to driving in favorable road and traffic conditions alone. Instead, the pattern of engaging in secondary tasks that develops when driving in partially-automated mode may be difficult to break and continue on unbridled even as driving and road conditions become more taxing. Our findings also add to the driver self-regulation literature. As the partially-automated system is engaged, it is hypothesized that this may lull drivers into a state of underload (McWilliams & Ward, 2021 ; Mishler & Chen, 2023 ), which may result in drivers becoming less capable of maintaining supervision over the functioning of the level-2 system. We argue that as drivers become aware of their state of underload, they start counteracting this pattern by engaging in activities that elevate their self-perceived workload. While our findings add to the literature on partial automation, our study has some limitations. First, drivers had no or little prior experience driving level-2 systems which might limit the validity of our findings to novel system users alone. Because we had no control over the duration and characteristics of the construction zone, the three zones were different in length. While we controlled for this factor in our analyses, it is possible that the pattern we observed might not extend to construction zones with different characteristics. Conclusions The literature on partially-automated driving has evidenced trends that are at odds with these systems’ purported safety. Our data add to this literature showing that the potential unintended consequences that operating level-2 systems has on driver behavior may extend to scenarios that present an even higher risk of vehicle-to-vehicle and vehicle-to-pedestrian collisions. While some automakers discourage or prohibit the use of their level-2 system in unfavorable road and traffic conditions, it’s worth noting that the decision of disengaging the system is oftentimes left up to the human driver. While the system may be aware of the ongoing road and traffic conditions through the information gathered through its sensors, we argue it should be engineered to automatically disengage itself knowing the higher safety risk being presented. While the existing research has just begun to learn more about how drivers use these systems in ideal road conditions, all but nothing is known about how level-2 automation behaves in scenarios representative of all real-world conditions. Declarations Author Contribution F.B. contributed to designing the study, data processing and analysis, and writing the manuscript. P.S. contributed to data collection and processing. N.J. contributed to data collection and processing. Acknowledgments The authors wish to acknowledge the generous support by the Ontario Ministry of Transportation, the Natural Science and Engineering Research Council and the Social Science and Humanities Research Council of Canada. Data Availability Data can be made available through a request to the corresponding author. References Banks, V. A., Eriksson, A., O’Donoghue, J., & Stanton, N. A. (2018). Is partially automated driving a bad idea? Observations from an on-road study. Applied Ergonomics, 68(October 2017), 138–145. https://doi.org/10.1016/j.apergo.2017.11.010 Biondi, F. N., & Jajo, N. (2024). On the impact of on-road partially-automated driving on drivers’ cognitive workload and attention allocation. Accident Analysis and Prevention, 200(March), 107537. https://doi.org/10.1016/j.aap.2024.107537 Biondi, F. N., Lohani, M., Hopman, R., Mills, S., Cooper, J. M., & Strayer, D. L. (2018). 80 MPH and out-of-the-loop : Effects of real-world semi-automated driving on driver workload and arousal. Proceedings of the Human Factors and Ergonomics Society Annual Meeting , 1878–1882. https://doi.org/https://doi.org/10.1177/1541931218621427 Biondi, F. N., McDonnell, A., Cooper, J., & Strayer, D. L. (2024). Using the ISO Detection response task to measure the cognitive load of driving four separate vehicles on two distinct highways. Transportation Research Part F: Traffic Psychology and Behaviour, 102(March), 260–269. https://doi.org/10.1016/j.trf.2024.02.013 Biondi, F. N., McDonnell, A. S., Mahmoodzadeh, M., Jajo, N., Balakumar Balasingam, & Strayer, D. L. (2023). Vigilance Decrement During On-Road Partially Automated Driving Across Four Systems. Human Factors. https://doi.org/10.1177/00187208231189658 Biondi, F., Turrill, J., Coleman, J. R., Cooper, J. M., & Strayer, D. L. (2015). Cognitive distraction impairs drivers’ anticipatory glances: an on-road study. Proceedings of the Eighth International Driving Symposium on Human Factors in Driver Assessment, Training and Vehicle Design Distractive, 23–29. Boehm, U., Matzke, D., Gretton, M., Castro, S., Cooper, J., Skinner, M., Strayer, D., & Heathcote, A. (2021). Real-time prediction of short-timescale fluctuations in cognitive workload. Cognitive Research: Principles and Implications, 6(1). https://doi.org/10.1186/s41235-021-00289-y Castro, S. C., Strayer, D. L., Matzke, D., & Heathcote, A. (2019). Cognitive workload measurement and modeling under divided attention. Journal of Experimental Psychology: Human Perception and Performance, 45(6), 826–839. https://doi.org/10.1037/xhp0000638 Dienes, Z. (2014). Using Bayes to get the most out of non-significant results. Frontiers in Psychology, 5(July), 1–17. https://doi.org/10.3389/fpsyg.2014.00781 Gaspar, J., & Carney, C. (2019). The Effect of Partial Automation on Driver Attention: A Naturalistic Driving Study. Human Factors, 61(8), 1261–1276. https://doi.org/10.1177/0018720819836310 ISO. (2015). Detection-response task (DRT) for assessing attentional effects of cognitive load in driving, ISO/DIS 17488. Khattak, A. J., Khattak, A. J., & Council, F. M. (2002). Effects of work zone presence on injury and non-injury crashes. Accident Analysis and Prevention, 34(1), 19–29. https://doi.org/10.1016/S0001-4575(00)00099-3 Lohani, M., Cooper, J. M., Erickson, G. G., Simmons, T. G., McDonnell, A. S., Carriero, A. E., Crabtree, K. W., & Strayer, D. L. (2021). No Difference in Arousal or Cognitive Demands Between Manual and Partially Automated Driving: A Multi-Method On-Road Study. Frontiers in Neuroscience, 15(June), 1–12. https://doi.org/10.3389/fnins.2021.577418 Lym, Y., & Chen, Z. (2021). Influence of built environment on the severity of vehicle crashes caused by distracted driving: A multi-state comparison. Accident Analysis and Prevention, 150(July 2020), 105920. https://doi.org/10.1016/j.aap.2020.105920 Ma, S., Hu, J., & Wang, R. (2023). Impact of Transition Areas on Driving Workload and Driving Behavior in Work Zones: A Naturalistic Driving Study. Applied Sciences (Switzerland), 13(21). https://doi.org/10.3390/app132111669 Mcdonnell, A. S., Crabtree, K. W., & City, S. L. (2023). This Is Your Brain on Autopilot 2.0: The Influence of Practice on Driver Workload and Engagement During On-Road, Partially Automated Driving Amy. Human Factors. https://doi.org/10.1177/00187208231201054 McDonnell, A. S., Simmons, T. G., Erickson, G. G., Lohani, M., Cooper, J. M., & Strayer, D. L. (2021). This Is Your Brain on Autopilot: Neural Indices of Driver Workload and Engagement During Partial Vehicle Automation. Human Factors. https://doi.org/10.1177/00187208211039091 McWilliams, T., & Ward, N. (2021). Underload on the Road: Measuring Vigilance Decrements During Partially Automated Driving. Frontiers in Psychology, 12(April), 1–13. https://doi.org/10.3389/fpsyg.2021.631364 Mishler, S., & Chen, J. (2023). Boring But Demanding: Using Secondary Tasks to Counter the Driver Vigilance Decrement for Partially Automated Driving. Human Factors. https://doi.org/10.1177/00187208231168697 Morando, A., Gershon, P., Mehler, B., & Reimer, B. (2021). A model for naturalistic glance behavior around Tesla Autopilot disengagements. Accident Analysis and Prevention, 161, 106348. https://doi.org/10.1016/j.aap.2021.106348 Naumann, R. B., Kreuger, L. K., Sandt, L., Lich, K. H., Bauchwitz, B., Kumfer1, W., & Combs, T. (2023). Examining the Safety Benefits of Partial Vehicle Automation Technologies in an Uncertain Future. https://aaafoundation.org/wp-content/uploads/2023/07/AAAFTS-Safety-Benefits-of-ADAS.pdf NHTSA. (2023). Part 573 Safety Recall Report 23V-085. 1–5. Noble, A. M., Miles, M., Perez, M. A., Guo, F., & Klauer, S. G. (2021). Evaluating driver eye glance behavior and secondary task engagement while using driving automation systems. Accident Analysis and Prevention, 151(March 2020), 105959. https://doi.org/10.1016/j.aap.2020.105959 NSC. (2024). Motor Vehicle Safety Issues - Work Zone. https://doi.org/10.1016/j.annemergmed.2016.04.045 NTSB. (2017). Collision Between a Car Operating With Automated Vehicle Control Systems and a Tractor-Semitrailer Truck Near Williston, Florida, May 7, 2016. NTSB. (2018). Collision Between a Sport Utility Vehicle Operating With Partial Driving Automation and a Crash Attenuator, Mountain View, California, March 23, 2018. NTSB. (2020). Tesla Crash Investigation Yields 9 NTSB Safety Recommendations (pp. 2017–2018). https://www.ntsb.gov/news/press-releases/Pages/NR20200225.aspx Quintana, D. S., & Williams, D. R. (2018). Bayesian alternatives for common null-hypothesis significance tests in psychiatry: A non-technical guide using JASP. BMC Psychiatry, 18(1), 1–8. https://doi.org/10.1186/s12888-018-1761-4 Racine, J. S. (2012). RStudio: A Platform-Independent IDE FOR R And Sweave. Journal of Applied Econometric, 27(1), 167–172. https://www.jstor.org/stable/41337225?seq=1#metadata_info_tab_contents Roshandeh, A. M., Zhou, B., & Behnood, A. (2016). Comparison of contributing factors in hit-and-run crashes with distracted and non-distracted drivers. Transportation Research Part F: Traffic Psychology and Behaviour, 38, 22–28. https://doi.org/10.1016/j.trf.2015.12.016 SAE. (2021). Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles. https://www.sae.org/standards/content/j3016_202104/ Shakouri, M., Ikuma, L. H., Aghazadeh, F., & Nahmens, I. (2018). Analysis of the sensitivity of heart rate variability and subjective workload measures in a driving simulator: The case of highway work zones. International Journal of Industrial Ergonomics, 66, 136–145. https://doi.org/10.1016/j.ergon.2018.02.015 Solís-Marcos, I., Galvao-Carmona, A., & Kircher, K. (2017). Reduced attention allocation during short periods of partially automated driving: An event-related potentials study. Frontiers in Human Neuroscience, 11(November), 1–13. https://doi.org/10.3389/fnhum.2017.00537 Yang, S., Kuo, J., & Lenné, M. G. (2021). Effects of Distraction in On-Road Level 2 Automated Driving: Impacts on Glance Behavior and Takeover Performance. Human Factors, 63(8), 1485–1497. https://doi.org/10.1177/0018720820936793 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 12 Mar, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 11 Oct, 2024 Reviews received at journal 10 Oct, 2024 Reviewers agreed at journal 07 Oct, 2024 Reviews received at journal 26 Sep, 2024 Reviewers agreed at journal 24 Sep, 2024 Reviews received at journal 26 Aug, 2024 Reviewers agreed at journal 15 Aug, 2024 Reviewers agreed at journal 15 Aug, 2024 Reviewers agreed at journal 15 Aug, 2024 Reviewers invited by journal 14 Aug, 2024 Editor assigned by journal 14 Aug, 2024 Editor invited by journal 12 Aug, 2024 Submission checks completed at journal 08 Jul, 2024 First submitted to journal 02 Jul, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4675940","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":325840657,"identity":"55f147fb-f4dd-4fb1-8049-a6ae3eeed762","order_by":0,"name":"Francesco Biondi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABDElEQVRIiWNgGAWjYBAC9gY4k4fhAAMDMw8/kHmAgQ23Fp4D6FokG0jRAgTMDAZgEXxapA8/e/Cjok6en4H34IGPO6xljG/kGB4uKKuTM29gfvgBmxa+NHPDnjOHDWc28CUcnHkmncfsRo7B4RnnDhvLHGAzlsCixZ6HwUyase1AgsEBHoPDvG2HgVpyNwAZBxJnAJ2KTQsPD/s3acZ/dQgtxjPAWupAWph/YNXCA7SlgRmhxUACrIUZpIUNuy08ZZI9x4B+aQb5pS2dR+LM+w+HeYB+kWBmM7PA7rBtEj9qgCHG3nv4w8c2a3v+9rTkzzzAEJNgb358A3dYM4BihLDIKBgFo2AUjAIiAQBQG1yVBLz5AwAAAABJRU5ErkJggg==","orcid":"","institution":"University of Windsor","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Francesco","middleName":"","lastName":"Biondi","suffix":""},{"id":325840658,"identity":"89ed666d-6d4e-45c6-9677-e0b3c90a786f","order_by":1,"name":"Praneet Sahoo","email":"","orcid":"","institution":"University of Windsor","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Praneet","middleName":"","lastName":"Sahoo","suffix":""},{"id":325840659,"identity":"702ae0d8-a1c3-4b7c-a907-dbfc3a802143","order_by":2,"name":"Noor Jajo","email":"","orcid":"","institution":"University of Windsor","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Noor","middleName":"","lastName":"Jajo","suffix":""}],"badges":[],"createdAt":"2024-07-02 17:56:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4675940/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4675940/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-93588-z","type":"published","date":"2025-03-12T15:58:51+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":62150306,"identity":"2724d56d-47d8-4b1f-a2c9-886cead3d034","added_by":"auto","created_at":"2024-08-09 20:28:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":803680,"visible":true,"origin":"","legend":"\u003cp\u003eMaps of the experimental route (A). The three zones are marked as follow: pre-construction in green, construction in orange, and post-construction in blue. (B) shows the road sign marking the beginning of the construction zone, and (C) shows the orange cone marking the end of the construction zone.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4675940/v1/bf872e178c6e3351f819a4c1.png"},{"id":62150305,"identity":"9610dc66-6155-4573-b93d-954bf2c7edf1","added_by":"auto","created_at":"2024-08-09 20:28:09","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":100806,"visible":true,"origin":"","legend":"\u003cp\u003ePercentage of time glancing at AOI by driving mode and construction zones. Error bars represent standard error.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4675940/v1/0564acca09acec89a7bbdab3.png"},{"id":78690136,"identity":"72fcd8ad-95e9-49d6-87a7-e93af0f511ca","added_by":"auto","created_at":"2025-03-17 16:14:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1416743,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4675940/v1/fdfc8a40-0c19-4181-a4cf-2f8082b8ae3f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Tesla Autopilot Through Constructions: Investigating the Effect of On-Road Partially-Automated Driving through Construction Zones","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePartially automated driving systems, also known as SAE level-2 automation, are systems capable of maintaining the lateral and longitudinal control of the vehicle in certain road and traffic scenarios (SAE, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Unlike lower or higher levels of automation wherein it is clear who is in control of the vehicle \u0026ndash; the human driver in SAE level-0 and 1, and the automated system in SAE level-4 and 5 \u0026ndash; in partial automation responsibilities are shared between the two. In particular, when the automated system is in control of the vehicle\u0026rsquo;s steering and accelerating, the human driver is still responsible for monitoring the system functioning and resuming control whenever necessary. The introduction of this technology has been promised to have a positive effect on safety especially in situations where the driver\u0026rsquo;s workload may be higher due to conditions like congested traffic or poorer visibility. Recent estimates point the adoption of this technology will be able to prevent 1 every 4 crashes and fatal injuries, and 1 every 3 fatalities in the year 2050 (Naumann et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDespite such rosy forecasts, the shared vehicle control between the human driver and the automated system is thought to be at the center of numerous fatal accidents. Recent investigations by US regulators point at driver complacency and misuse as contributing factors to these accidents. For example, the National Transportation Safety Board (NTSB) investigation on the 2016 fatal accident, wherein a partially-automated vehicle crashed into a tractor semitrailer in Florida, reports that the absence of design constraints that limited the use of these systems only when appropriate could give rise to drivers misusing these vehicles (NTSB, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In a report following a similar crash in 2018, NTSB also points at the vehicle\u0026rsquo;s ineffective monitoring of driver engagement, which resulted in driver inattentiveness and complacency, as a probable cause of the accident (NTSB, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The 2023 NHTSA and Transport Canada recall on Tesla\u0026rsquo;s Autosteer system also indicate that, because of its design, drivers may misuse the partially automated system which could lead to an increased risk of collisions (NHTSA, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThe conclusions reached in these investigations (also see NTSB, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) find support in the broader literature on partial automation use. Studies largely point at the higher risk of driver disengagement and distraction occurring whenever the automated system is engaged. Prior work by Biondi et al. (Biondi et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Biondi \u0026amp; Jajo, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) investigated differences in drivers\u0026rsquo; workload and attention allocation when transitioning from manual to partially-automated driving. While mental workload seemed largely unaffected by driving mode, drivers became less attentive toward the driving scene as the control of the vehicle was handed over to the automated system. When compared to manual driving, such attention decline appeared to peak over time during partial automation, an effect that seemed consistent across distinct level-2 systems (Biondi et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Findings by Morando et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and Yang et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) with drivers operating a Tesla vehicle in Autopilot mode point at similar patterns. In the time preceding the transition to manual driving, Morando et al. observed a higher frequency of longer glances directed toward the vehicle\u0026rsquo;s center-stack. The study by Yang et al. also highlights that lesser engagement in the driving task when the Autopilot system is on degraded the driver\u0026rsquo;s ability to resume manual control of the vehicle post-transition. Naturalistic data by Noble et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) align with this literature showing a higher prevalence of cellphone use during partially-automated driving.\u003c/p\u003e \u003cp\u003eThese findings warn caution about the use of these systems, especially in conditions where the risks of vehicle-to-pedestrian or vehicle-to-vehicle collisions are heightened. Construction zones represent particularly perilous road sections. In 2022, 891 people were killed and 37,701 were injured in construction zone crashes in the US alone (NSC, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Work by Khattak et al. (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) shows that the risk of crashes during construction zones is 21.5% higher than that in pre-construction zones, with the formers also exhibiting a 23.8% and 17.3% increase in non-injury and injury crash risks, respectively. A more recent analysis conducted by Lym and Chen (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) indicate that the severity of vehicle crashes resulting from driver distraction tends to be higher in construction zones. Consistent findings are offered by Roshandeh et al. (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) who also observed an increased risk of hit-and-run crashes in construction zones in the presence of a distracted driver.\u003c/p\u003e \u003cp\u003eWith construction zones being so dangerous, the adoption of partially automated systems holds the potential of benefiting safety in these collision-prone road sections. Yet little evidence is available on the effect that using these systems has on driver behavior when driving in construction zones. Related literature has in fact largely focused on investigating driver interaction with these systems in favorable traffic and road conditions (Banks et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Gaspar \u0026amp; Carney, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Yang et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). With this in mind, the current study aims to fill this gap by having participants drive a Tesla vehicle in either partially automated or manual mode in three distinct zones: pre-construction, construction, and post-construction. The first objective is to investigate the combined effect that driving mode and zone has on cognitive workload. Prior research has shown mixed findings about the effect of level-2 automated driving on cognitive workload with studies either showing no differences between partial automation and manual driving (Lohani et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Mcdonnell et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) or a decrease in cognitive workload in level-2 mode (Biondi et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Sol\u0026iacute;s-Marcos et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). With the driving task arguably becoming more demanding during construction zones (Ma et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Shakouri et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), we expect cognitive workload to increase when transitioning from the pre-construction zone into the construction zone in both manual and level-2 mode. The second objective is to investigate the combined effect of driving mode and zone on glance allocation. Prior research has shown glance allocation toward driving-unrelated areas to increase during level-2 driving (Morando et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Yang et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In the current study we also expect a similar pattern during level-2 driving, but with glances toward the forward roadway possibly increasing with the beginning of the construction zone.\u003c/p\u003e "},{"header":"Method","content":" \u003cp\u003eParticipants\u003c/p\u003e \u003cp\u003eTwenty-one volunteers were recruited from the University of Windsor. Their average age and standard deviation of age were 22 years and 4.36 years, respectively. All participants were fluent English speakers, had normal or corrected-to-normal vision and hearing, held a valid driver\u0026rsquo;s license, had proof of car insurance, and had not been the at-fault driver in an accident within the past 2 years. Participants were also required to complete 30-minute defensive driving course and be affiliated with the University of Windsor. A University of Windsor Research Ethics Board approval (REB #20\u0026ndash;141) was obtained for the study. All methods were performed in accordance with the relevant guidelines and regulations.\u003c/p\u003e \u003cp\u003eExperimental Design\u003c/p\u003e \u003cp\u003eA factorial design with two independent variables were used in this study. The within-subject independent variables were driving mode (2 levels) and zones (3 levels). Participants drove a vehicle in either manual or level-2 automation mode, and through three zones: pre-construction, construction, and post-construction. Dependent measures included the performance to the ISO Detection Response Task (ISO, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) and the percentage of time spent looking at each of the following areas of interest (AOI): forward roadway, vehicle\u0026rsquo;s touchscreen, side mirrors, and rearview mirror.\u003c/p\u003e \u003cp\u003eEquipment and Data Processing\u003c/p\u003e\n\u003ch3\u003eVehicle\u003c/h3\u003e\n\u003cp\u003eA 2022 Tesla Model 3 was used for the study. Participants drove the vehicle in manual and level-2 mode. In level-2 mode, the vehicle was driven with both Adaptive Cruise Control (ACC) and Lane Keeping Assist System (LKAS) engaged. ACC maintained the vehicle at a set speed and distance from the vehicle in front and, during the study, it was set to the maximum distance of 7 car lengths. LKAS maintained the vehicle within the lane and, during the study, drivers were instructed to occupy the right lane when two lanes were available or the middle lane when three lanes were available. This was done to minimize lane changing and the need to overtake slower traveling vehicles. When working together, ACC and LKAS meet all requirements necessary of SAE level-2 automation.In manual mode, participants drove the vehicle manually with no assistance from either ACC or LKAS.\u003c/p\u003e\n\u003ch3\u003eRoute\u003c/h3\u003e\n\u003cp\u003eTwo routes were considered for the study. A familiarization route was used for participants to familiarize with the vehicle and the level-2 system. Participants drove a loop through Huron Church road and Ontario Highway 3 between approximately the University of Windsor campus and the St. Clair College campus in Windsor, ON. The experimental route consisted of a section of Ontario Highway 401 between Exit 13 in Windsor, ON and Exit 81 near Chatham, ON. The road consists of a combination of three-lane and two-lane dual carriageways that provides sufficient space between vehicles. The daily traffic volume on the route was approximately 25,000 vehicles. For the current study, three distinct zones were identified: pre-construction, construction, and post-construction (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). The pre-construction zone consisted of the road section from the start of the drive (kilometer 13) to the beginning of the construction zone (kilometer 50). The start and end of the construction zone were marked by a road sign (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB) and the presence of the last orange cone (kilometer 65; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC), respectively. The post-construction zone followed until the end of the drive at kilometer 81.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e\u003c/h2\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eDetection Response Task\u003c/h2\u003e \u003cp\u003eThe vibrotactile version of the Detection Response Task (DRT) manufactured by Red Scientific Ltd (Salt Lake City, UT, USA) was used in the study. A vibrotactile motor was placed on the inside of the participants\u0026rsquo; left elbow area and a microswitch was attached to either the index or middle finger of the left hand. A vibration, similar to that emitted by a cellphone, was presented every 3 to 5 seconds with a duration of 1 second. Participants were instructed to respond to the vibration by pressing a microswitch as fast as possible. Reaction times (RT) in milliseconds and hit rates were recorded. RT were recorded as the time interval between the onset of the vibrotactile stimulus and the depression of the microswitch. RT faster than 100 ms or longer than 2,500 ms were eliminated from the calculation. Nonresponses or responses produced later than 2,500 ms were considered as misses. Average RTs were calculated by driving mode and zone.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eCameras\u003c/h2\u003e \u003cp\u003eThe vehicle was equipped with three GoPro HERO8 Black cameras. The driver view camera was points at the driver and footage from this camera was used to record eye glances and eye movements. The front view camera was pointed at the forward roadway. The touchscreen view camera was pointed at the vehicle\u0026rsquo;s touchscreen. The cameras recorded at 1080p at 240 frames per second.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eGlances\u003c/h2\u003e \u003cp\u003eFootage from the driver camera was manually coded by two trained coders. Four areas of interest (AOI) were identified: forward roadway, vehicle\u0026rsquo;s touchscreen, side mirrors, and rearview mirror. Glances directed toward the forward roadway included any glance directed toward the center, the left, or the right side of the road to, e.g., inspect the road or vehicle in front, or check for potential hazards or signs. Glances directed at the vehicle\u0026rsquo;s touchscreen included any glance directed at the touchscreen displaying information about the vehicle\u0026rsquo;s functioning. Glances directed at the side mirrors and rearview mirrors included any glance directed toward these areas to, e.g., check for the surrounding vehicles. The selection of these AOI is consistent with prior work by Biondi et al. (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) and Gaspar and Carney (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Two independent coders were assigned to each video and manually coded it frame-by-frame to measure the time each driver spent looking at each AOI during the manual and L2 drive. Possible discrepancies between the two coders were flagged and reviewed by a third coder. Cohen\u0026rsquo;s kappa was calculated as a measure of inter-coder reliability. A Cohen\u0026rsquo;s kappa of 0.85 was found indicating high consistency between the two coders (McHugh, 2012). The percentage of time spent looking at each AOI was calculated as the ratio between the time spent looking at each AOI divided by the total time spent in each of the three zones (pre-construction, construction, post-construction). For example, if a driver spent 30 seconds looking at the forward roadway while driving in a construction zone that took a total of 60 seconds to travel, the percentage of time spent looking at that AOI in that zone would be 50%. This metric was calculated as such to account for differences in zone lengths and the time it took participants to travel through them. Average percentages of time looking at each AOI were calculated by driving mode and zone.\u003c/p\u003e \u003cp\u003eProcedure\u003c/p\u003e \u003cp\u003e\u003cem\u003eStudy intake\u003c/em\u003e. Prior to the day of the study, participants were provided with a video tutorial on how to use the level-2 system on the vehicle. This was done to ensure that all participants were provided consistent information about how to use the system. On the day of the study, participants met with the research associate and were directed to where the vehicle was parked. When inside the vehicle, participants were asked to review the signed informed consent form and REB checklist and ask any questions they may have. The research associate then advised them to sit in the car to learn how to adjust the mirrors, seats, and steering wheel.\u003c/p\u003e \u003cp\u003e \u003cem\u003eFamiliarization phase\u003c/em\u003e. During this phase, participants were given 1 minute to practice completing the DRT when the vehicle was stationary. They were also reminded to put their phones on mute, and that the could not make use of their phones for the entire duration of the drive, unless in case of an emergency. Participants drove the familiarization route in manual and L2 mode for as long as they needed, which took up to 20 min.\u003c/p\u003e \u003cp\u003e\u003cem\u003eExperimental phase\u003c/em\u003e. Once participants verbally confirmed they were comfortable driving the vehicle in manual and L2 mode, the experimental phase begun. Participants drove the experimental route twice: once in level-2 mode and once in manual mode. The order of the two drives was counterbalanced across participants. Each drive took approximately 40 minutes to complete. Participants exited the highway at kilometer 81 near Chatham, ON and parked at a gas station where they could take a 15-minute break. After the break, they reentered the highway in the opposite direction and the second drive begun. The second drive ended at exit 13 in Windsor, ON at which point participants were instructed to drive back to the University of Windsor campus. The experimental phase took up to 2 hours.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003eBayes factor analyses were adopted for data analyses. Unlike the traditional null-hypothesis statistical testing (NHST) which relies on the p-value to determine whether the null hypothesis is accepted, Bayesian analyses set up two competing models, one for the null hypothesis and one for the alternative hypothesis, and estimate which of the two models is more likely to generate the data at hand. Bayesian analyses transform p-values into direct evidence against the null hypotheses (Held and Ott, 2018). A Bayes Factor (BF) is calculated as the ratio between the marginal likelihood of the null model and that of the alternative model (Quintana \u0026amp; Williams, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). A BF equal to X indicates that the data is X times more likely under the alternative hypotheses than under the null hypothesis. BF range between 0 and infinity. BF less than 0.33 indicate evidence in support of the null hypothesis, with smaller values suggesting stronger evidence for H\u003csub\u003e0\u003c/sub\u003e. BF greater than 3 indicate evidence in support of the alternative hypothesis, with larger values suggesting stronger evidence for H\u003csub\u003e1\u003c/sub\u003e (Dienes, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Unlike NHST, Bayesian analyses provide evidence in support of either the null or the alternative hypothesis, as well as offer information about the strength of the evidence. Data processing and analyses were conducted using R (version 4.1.0) and RStudio (version 2023.03.0 (Racine, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2012\u003c/span\u003e)). The tidyverse (version 2.0) and BayesFactor (version 9.12) libraries were adopted for data processing and Bayesian analyses, respectively.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eResults are presented by objectives.\u003c/p\u003e \u003cp\u003eInvestigate the combined effect that driving mode and zone has on cognitive workload\u003c/p\u003e \u003cp\u003eA model was set up with driving mode (2 levels: manual, level-2) and zone (3 levels: pre-construction, construction, post-construction) as the independent factors and DRT RT as the dependent measure. A BF of 0.19 was found for driving mode indicating strong evidence in favor of the null hypothesis. This finding aligns with prior findings by McDonnell et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) who also found no significant changes in drivers\u0026rsquo; cognitive workload between driving in manual mode and driving with the level-2 system engaged. Strong evidence in support of the null hypothesis was also found for the main effect of zone (BF\u0026thinsp;=\u0026thinsp;0.18) indicating that, although DRT RT increased from the pre-construction to the construction zone, this difference was nonsignificant. A BF of 0.005 was found for the zone by mode interaction indicating no significant interaction between these two factors. DRT RTs are presented in Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMean and standard error (SE) of DRT RT in milliseconds by zone and driving mode.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"10\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"8\" nameend=\"c10\" namest=\"c3\"\u003e \u003cp\u003eZone\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003ePre-construction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003eConstruction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c10\" namest=\"c9\"\u003e \u003cp\u003ePost-construction\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSE\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eManual\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e526\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e41.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e622\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e71.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e602\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e46.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLevel-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e549\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e594\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e52.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e603\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e48.8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eInvestigate the combined effect of driving mode and zone on glance allocation\u003c/p\u003e \u003cp\u003eModels were set up with driving mode (2 levels: manual, level-2), zone (3 levels: pre-construction, construction, post-construction) and AOI (4 levels: forward roadway, side mirrors, rearview mirror, touchscreen) as the independent factors and the percentage of time spent looking at AOIs as the dependent measure. A BF of 2.95 x 10\u003csup\u003e642\u003c/sup\u003e was found for AOIs indicating differences in the time spent looking at each AOI. Based on this result, separate models were set up, each investigating separate AOIs. For forward roadway glances, a BF of 5.4 x 10\u003csup\u003e3\u003c/sup\u003e was found for mode indicating a strong difference between the two modes. Whereas drivers spent an average of 96.2% (SE\u0026thinsp;=\u0026thinsp;0.83%) of the time looking at the forward roadway during manual mode, this decreased to 91% (SE\u0026thinsp;=\u0026thinsp;0.44%) during level-2 driving. A BF of 0.08 was found for zone suggesting that the time spent looking at the forward roadway did not change across zones (pre-construction: M\u0026thinsp;=\u0026thinsp;94.3%, SE\u0026thinsp;=\u0026thinsp;0.64%; construction: M\u0026thinsp;=\u0026thinsp;93.9%, SE\u0026thinsp;=\u0026thinsp;1.01%, post-construction\u0026thinsp;=\u0026thinsp;93.5%, SE\u0026thinsp;=\u0026thinsp;0.98%). Similar analyses were conducted for touchscreen. A BF of 4.2 x 10\u003csup\u003e3\u003c/sup\u003e was found for mode indicating that drivers spent more time looking at the vehicle\u0026rsquo;s touchscreen during level-2 mode (M\u0026thinsp;=\u0026thinsp;6.14%, SE\u0026thinsp;=\u0026thinsp;0.61%) relative to manual mode (M\u0026thinsp;=\u0026thinsp;2.79%, SE\u0026thinsp;=\u0026thinsp;0.35%). A BF\u0026thinsp;=\u0026thinsp;0.07 was found for zone suggesting that the time spent looking at the touchscreen did not change across zones (pre-construction: M\u0026thinsp;=\u0026thinsp;4.36%, SE\u0026thinsp;=\u0026thinsp;0.58%; construction: M\u0026thinsp;=\u0026thinsp;4.26%, SE\u0026thinsp;=\u0026thinsp;0.68%, post-construction\u0026thinsp;=\u0026thinsp;4.73%, SE\u0026thinsp;=\u0026thinsp;0.74%). Analyses conducted for glances directed at the side mirrors revealed a BF of 1.81 for the main effect of mode and a BF of 0.11 for zone suggesting that the percentage of time looking at this AOI did not change across modes nor zones. Similar results were found for the rearview mirror. A BF of 0.64 was found for mode and a BF of 0.08 was found for zone, suggesting no differences in the time spent looking at this AOI by mode or zone. Data for forward roadway and touchscreen AOIs are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study we set out to investigate two distinct objectives. The first objective was to investigate the effect that driving mode and zone had on drivers\u0026rsquo; cognitive workload. Prior research on the topic has produced mixed findings. Studies by McDonnell et al. (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and Lohani et al. (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) showed no differences in drivers\u0026rsquo; workload between the two modes. Conflicting patterns were observed by Biondi et al. (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) and Solis-Marco et al. (2017) who found lower mental workload when driving in level-2 mode. Our results align with the work by McDonnell et al. and Lohani et al. with no differences being found in DRT RT between manual and partial automation. No differences in DRT RT were found between the three zones under consideration indicating that cognitive workload did not change across the pre-construction, construction, and post-construction zones. This result is at odds with prior research showing an increase in driving demand when travelling through construction zones (Ma et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Shakouri et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Although the ISO DRT has proven sensitive in discriminating changes in cognitive workload induced by diverse driving and road conditions (Biondi et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Boehm et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Castro et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), it is still possible that the changes in driving demand elicited by our three zones may have not been sufficient to produce significant changes in DRT performance.\u003c/p\u003e \u003cp\u003eOur second objective was to investigate the effect of driving mode and zone on glance allocation. Analyses conducted on glances directed at the forward roadway evidenced a decline in the percentage of time looking at this AOI during level-2 driving. This datum aligns with prior findings by Morando et al. (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and Yang et al. (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) evidencing a decline in visual attention toward the forward roadway when the partially automated system is engaged. As the time spent looking at the forward roadway decreased, the time spent looking at the vehicle\u0026rsquo;s touchscreen increased in level-2 mode. Again, this pattern is consistent with prior work by Noble et al. (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) who also suggest that driving in partially automated mode increases the likelihood of engaging in driving-unrelated activities. Analyses investigating changes in glance allocation by zone revealed important patterns. No differences in the time spent looking at the forward roadway was found between pre-construction, construction, and post-construction zones, indicating that the presence of the construction zone didn\u0026rsquo;t bring drivers to increase the time spent looking at the forward roadway.\u003c/p\u003e \u003cp\u003eThis finding highlights a dangerous pattern. It was indeed expected that as drivers transitioned into the construction zone, this may have increased the amount of forward glances executed to anticipate the presence of hazards and construction workers. The fact that, during the construction zone, drivers kept glancing away from the road toward the vehicle\u0026rsquo;s touchscreen to the same extend they did in the pre-construction zone suggests that the potential for distraction resulting from level-2 system (see Noble et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) might not be limited to driving in favorable road and traffic conditions alone. Instead, the pattern of engaging in secondary tasks that develops when driving in partially-automated mode may be difficult to break and continue on unbridled even as driving and road conditions become more taxing. Our findings also add to the driver self-regulation literature. As the partially-automated system is engaged, it is hypothesized that this may lull drivers into a state of underload (McWilliams \u0026amp; Ward, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Mishler \u0026amp; Chen, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), which may result in drivers becoming less capable of maintaining supervision over the functioning of the level-2 system. We argue that as drivers become aware of their state of underload, they start counteracting this pattern by engaging in activities that elevate their self-perceived workload.\u003c/p\u003e \u003cp\u003eWhile our findings add to the literature on partial automation, our study has some limitations. First, drivers had no or little prior experience driving level-2 systems which might limit the validity of our findings to novel system users alone. Because we had no control over the duration and characteristics of the construction zone, the three zones were different in length. While we controlled for this factor in our analyses, it is possible that the pattern we observed might not extend to construction zones with different characteristics.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe literature on partially-automated driving has evidenced trends that are at odds with these systems\u0026rsquo; purported safety. Our data add to this literature showing that the potential unintended consequences that operating level-2 systems has on driver behavior may extend to scenarios that present an even higher risk of vehicle-to-vehicle and vehicle-to-pedestrian collisions. While some automakers discourage or prohibit the use of their level-2 system in unfavorable road and traffic conditions, it\u0026rsquo;s worth noting that the decision of disengaging the system is oftentimes left up to the human driver. While the system may be aware of the ongoing road and traffic conditions through the information gathered through its sensors, we argue it should be engineered to automatically disengage itself knowing the higher safety risk being presented. While the existing research has just begun to learn more about how drivers use these systems in ideal road conditions, all but nothing is known about how level-2 automation behaves in scenarios representative of all real-world conditions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eF.B. contributed to designing the study, data processing and analysis, and writing the manuscript. P.S. contributed to data collection and processing. N.J. contributed to data collection and processing.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eThe authors wish to acknowledge the generous support by the Ontario Ministry of Transportation, the Natural Science and Engineering Research Council and the Social Science and Humanities Research Council of Canada.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eData can be made available through a request to the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBanks, V. A., Eriksson, A., O\u0026rsquo;Donoghue, J., \u0026amp; Stanton, N. A. (2018). Is partially automated driving a bad idea? Observations from an on-road study. Applied Ergonomics, 68(October 2017), 138\u0026ndash;145. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.apergo.2017.11.010\u003c/span\u003e\u003cspan address=\"10.1016/j.apergo.2017.11.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiondi, F. N., \u0026amp; Jajo, N. (2024). On the impact of on-road partially-automated driving on drivers\u0026rsquo; cognitive workload and attention allocation. Accident Analysis and Prevention, 200(March), 107537. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.aap.2024.107537\u003c/span\u003e\u003cspan address=\"10.1016/j.aap.2024.107537\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiondi, F. N., Lohani, M., Hopman, R., Mills, S., Cooper, J. M., \u0026amp; Strayer, D. L. (2018). 80 MPH and out-of-the-loop : Effects of real-world semi-automated driving on driver workload and arousal. \u003cem\u003eProceedings of the Human Factors and Ergonomics Society Annual Meeting\u003c/em\u003e, 1878\u0026ndash;1882. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1177/1541931218621427\u003c/span\u003e\u003cspan address=\"10.1177/1541931218621427\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiondi, F. N., McDonnell, A., Cooper, J., \u0026amp; Strayer, D. L. (2024). Using the ISO Detection response task to measure the cognitive load of driving four separate vehicles on two distinct highways. Transportation Research Part F: Traffic Psychology and Behaviour, 102(March), 260\u0026ndash;269. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.trf.2024.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.trf.2024.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiondi, F. N., McDonnell, A. S., Mahmoodzadeh, M., Jajo, N., Balakumar Balasingam, \u0026amp; Strayer, D. L. (2023). Vigilance Decrement During On-Road Partially Automated Driving Across Four Systems. Human Factors. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00187208231189658\u003c/span\u003e\u003cspan address=\"10.1177/00187208231189658\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBiondi, F., Turrill, J., Coleman, J. R., Cooper, J. M., \u0026amp; Strayer, D. L. (2015). Cognitive distraction impairs drivers\u0026rsquo; anticipatory glances: an on-road study. Proceedings of the Eighth International Driving Symposium on Human Factors in Driver Assessment, Training and Vehicle Design Distractive, 23\u0026ndash;29.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoehm, U., Matzke, D., Gretton, M., Castro, S., Cooper, J., Skinner, M., Strayer, D., \u0026amp; Heathcote, A. (2021). Real-time prediction of short-timescale fluctuations in cognitive workload. Cognitive Research: Principles and Implications, 6(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s41235-021-00289-y\u003c/span\u003e\u003cspan address=\"10.1186/s41235-021-00289-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCastro, S. C., Strayer, D. L., Matzke, D., \u0026amp; Heathcote, A. (2019). Cognitive workload measurement and modeling under divided attention. Journal of Experimental Psychology: Human Perception and Performance, 45(6), 826\u0026ndash;839. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/xhp0000638\u003c/span\u003e\u003cspan address=\"10.1037/xhp0000638\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDienes, Z. (2014). Using Bayes to get the most out of non-significant results. Frontiers in Psychology, 5(July), 1\u0026ndash;17. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpsyg.2014.00781\u003c/span\u003e\u003cspan address=\"10.3389/fpsyg.2014.00781\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGaspar, J., \u0026amp; Carney, C. (2019). The Effect of Partial Automation on Driver Attention: A Naturalistic Driving Study. Human Factors, 61(8), 1261\u0026ndash;1276. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0018720819836310\u003c/span\u003e\u003cspan address=\"10.1177/0018720819836310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eISO. (2015). Detection-response task (DRT) for assessing attentional effects of cognitive load in driving, ISO/DIS 17488.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhattak, A. J., Khattak, A. J., \u0026amp; Council, F. M. (2002). Effects of work zone presence on injury and non-injury crashes. Accident Analysis and Prevention, 34(1), 19\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0001-4575(00)00099-3\u003c/span\u003e\u003cspan address=\"10.1016/S0001-4575(00)00099-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLohani, M., Cooper, J. M., Erickson, G. G., Simmons, T. G., McDonnell, A. S., Carriero, A. E., Crabtree, K. W., \u0026amp; Strayer, D. L. (2021). No Difference in Arousal or Cognitive Demands Between Manual and Partially Automated Driving: A Multi-Method On-Road Study. Frontiers in Neuroscience, 15(June), 1\u0026ndash;12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnins.2021.577418\u003c/span\u003e\u003cspan address=\"10.3389/fnins.2021.577418\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLym, Y., \u0026amp; Chen, Z. (2021). Influence of built environment on the severity of vehicle crashes caused by distracted driving: A multi-state comparison. Accident Analysis and Prevention, 150(July 2020), 105920. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.aap.2020.105920\u003c/span\u003e\u003cspan address=\"10.1016/j.aap.2020.105920\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa, S., Hu, J., \u0026amp; Wang, R. (2023). Impact of Transition Areas on Driving Workload and Driving Behavior in Work Zones: A Naturalistic Driving Study. Applied Sciences (Switzerland), 13(21). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/app132111669\u003c/span\u003e\u003cspan address=\"10.3390/app132111669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcdonnell, A. S., Crabtree, K. W., \u0026amp; City, S. L. (2023). This Is Your Brain on Autopilot 2.0: The Influence of Practice on Driver Workload and Engagement During On-Road, Partially Automated Driving Amy. Human Factors. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00187208231201054\u003c/span\u003e\u003cspan address=\"10.1177/00187208231201054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcDonnell, A. S., Simmons, T. G., Erickson, G. G., Lohani, M., Cooper, J. M., \u0026amp; Strayer, D. L. (2021). This Is Your Brain on Autopilot: Neural Indices of Driver Workload and Engagement During Partial Vehicle Automation. Human Factors. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00187208211039091\u003c/span\u003e\u003cspan address=\"10.1177/00187208211039091\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcWilliams, T., \u0026amp; Ward, N. (2021). Underload on the Road: Measuring Vigilance Decrements During Partially Automated Driving. Frontiers in Psychology, 12(April), 1\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpsyg.2021.631364\u003c/span\u003e\u003cspan address=\"10.3389/fpsyg.2021.631364\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMishler, S., \u0026amp; Chen, J. (2023). Boring But Demanding: Using Secondary Tasks to Counter the Driver Vigilance Decrement for Partially Automated Driving. Human Factors. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/00187208231168697\u003c/span\u003e\u003cspan address=\"10.1177/00187208231168697\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorando, A., Gershon, P., Mehler, B., \u0026amp; Reimer, B. (2021). A model for naturalistic glance behavior around Tesla Autopilot disengagements. Accident Analysis and Prevention, 161, 106348. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.aap.2021.106348\u003c/span\u003e\u003cspan address=\"10.1016/j.aap.2021.106348\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaumann, R. B., Kreuger, L. K., Sandt, L., Lich, K. H., Bauchwitz, B., Kumfer1, W., \u0026amp; Combs, T. (2023). Examining the Safety Benefits of Partial Vehicle Automation Technologies in an Uncertain Future. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://aaafoundation.org/wp-content/uploads/2023/07/AAAFTS-Safety-Benefits-of-ADAS.pdf\u003c/span\u003e\u003cspan address=\"https://aaafoundation.org/wp-content/uploads/2023/07/AAAFTS-Safety-Benefits-of-ADAS.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNHTSA. (2023). Part 573 Safety Recall Report 23V-085. 1\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoble, A. M., Miles, M., Perez, M. A., Guo, F., \u0026amp; Klauer, S. G. (2021). Evaluating driver eye glance behavior and secondary task engagement while using driving automation systems. Accident Analysis and Prevention, 151(March 2020), 105959. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.aap.2020.105959\u003c/span\u003e\u003cspan address=\"10.1016/j.aap.2020.105959\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNSC. (2024). Motor Vehicle Safety Issues - Work Zone. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.annemergmed.2016.04.045\u003c/span\u003e\u003cspan address=\"10.1016/j.annemergmed.2016.04.045\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNTSB. (2017). Collision Between a Car Operating With Automated Vehicle Control Systems and a Tractor-Semitrailer Truck Near Williston, Florida, May 7, 2016.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNTSB. (2018). Collision Between a Sport Utility Vehicle Operating With Partial Driving Automation and a Crash Attenuator, Mountain View, California, March 23, 2018.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNTSB. (2020). Tesla Crash Investigation Yields 9 NTSB Safety Recommendations (pp. 2017\u0026ndash;2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ntsb.gov/news/press-releases/Pages/NR20200225.aspx\u003c/span\u003e\u003cspan address=\"https://www.ntsb.gov/news/press-releases/Pages/NR20200225.aspx\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuintana, D. S., \u0026amp; Williams, D. R. (2018). Bayesian alternatives for common null-hypothesis significance tests in psychiatry: A non-technical guide using JASP. BMC Psychiatry, 18(1), 1\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12888-018-1761-4\u003c/span\u003e\u003cspan address=\"10.1186/s12888-018-1761-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacine, J. S. (2012). RStudio: A Platform-Independent IDE FOR R And Sweave. Journal of Applied Econometric, 27(1), 167\u0026ndash;172. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jstor.org/stable/41337225?seq=1#metadata_info_tab_contents\u003c/span\u003e\u003cspan address=\"https://www.jstor.org/stable/41337225?seq=1#metadata_info_tab_contents\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoshandeh, A. M., Zhou, B., \u0026amp; Behnood, A. (2016). Comparison of contributing factors in hit-and-run crashes with distracted and non-distracted drivers. Transportation Research Part F: Traffic Psychology and Behaviour, 38, 22\u0026ndash;28. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.trf.2015.12.016\u003c/span\u003e\u003cspan address=\"10.1016/j.trf.2015.12.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSAE. (2021). Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.sae.org/standards/content/j3016_202104/\u003c/span\u003e\u003cspan address=\"https://www.sae.org/standards/content/j3016_202104/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShakouri, M., Ikuma, L. H., Aghazadeh, F., \u0026amp; Nahmens, I. (2018). Analysis of the sensitivity of heart rate variability and subjective workload measures in a driving simulator: The case of highway work zones. International Journal of Industrial Ergonomics, 66, 136\u0026ndash;145. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ergon.2018.02.015\u003c/span\u003e\u003cspan address=\"10.1016/j.ergon.2018.02.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSol\u0026iacute;s-Marcos, I., Galvao-Carmona, A., \u0026amp; Kircher, K. (2017). Reduced attention allocation during short periods of partially automated driving: An event-related potentials study. Frontiers in Human Neuroscience, 11(November), 1\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fnhum.2017.00537\u003c/span\u003e\u003cspan address=\"10.3389/fnhum.2017.00537\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, S., Kuo, J., \u0026amp; Lenn\u0026eacute;, M. G. (2021). Effects of Distraction in On-Road Level 2 Automated Driving: Impacts on Glance Behavior and Takeover Performance. Human Factors, 63(8), 1485\u0026ndash;1497. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0018720820936793\u003c/span\u003e\u003cspan address=\"10.1177/0018720820936793\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Tesla Autopilot, Cognitive Workload, Glances, Road Safety, Partial Automation","lastPublishedDoi":"10.21203/rs.3.rs-4675940/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4675940/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePartially-automated driving systems are designed to control the vehicle\u0026rsquo;s speed and acceleration without input from the human driver on the condition that the driver maintains alertness. These systems are promised to make driving safer especially when driving in road sections exhibiting a higher risk of collisions like construction zones. Despite this, little knowledge is available on how these systems are used in these accident-prone areas and the effect they may have on drivers\u0026rsquo; workload and glance allocation. This study aims to fill this gap by having participants drive a Tesla vehicle in Autopilot and manual mode through three road sections: pre-construction, construction, and post-construction. Results show no differences in cognitive workload by driving mode or construction zone. An increase in glances directed away from the forward roadway toward the vehicle\u0026rsquo;s touchscreen was observed during partially-automated driving in the pre-construction zone, a pattern that, notably, continued on when driving throughout the construction zone. These findings adds to the literature on the human factors of partial automation. More importantly, because drivers failed to increase the amount of time looking at the forward roadway when entering the construction zone, they show the perniciousness of partially-automated driving and the detrimental effect these systems may have on safety.\u003c/p\u003e","manuscriptTitle":"Tesla Autopilot Through Constructions: Investigating the Effect of On-Road Partially-Automated Driving through Construction Zones","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-09 20:28:04","doi":"10.21203/rs.3.rs-4675940/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-11T09:44:20+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-10T20:45:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"239993172811909003108629477358794723300","date":"2024-10-07T22:26:30+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-09-26T12:55:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"335827252649386700703630680671324272843","date":"2024-09-25T00:06:10+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-27T03:21:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"265751818901363226437942900349286635897","date":"2024-08-15T17:49:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"149414180206667092733317459634346524885","date":"2024-08-15T15:31:03+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"179725887869176847319054305557667355719","date":"2024-08-15T12:38:51+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-08-15T02:12:21+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-15T00:39:30+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-08-12T19:07:21+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-07-09T03:58:33+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-07-02T17:54:55+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8616ec6d-48f1-4811-9828-9c6d0b833d0d","owner":[],"postedDate":"August 9th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":34476024,"name":"Biological sciences/Psychology/Human behaviour"},{"id":34476025,"name":"Earth and environmental sciences/Environmental social sciences/Psychology and behaviour"}],"tags":[],"updatedAt":"2025-03-17T16:12:49+00:00","versionOfRecord":{"articleIdentity":"rs-4675940","link":"https://doi.org/10.1038/s41598-025-93588-z","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-03-12 15:58:51","publishedOnDateReadable":"March 12th, 2025"},"versionCreatedAt":"2024-08-09 20:28:04","video":"","vorDoi":"10.1038/s41598-025-93588-z","vorDoiUrl":"https://doi.org/10.1038/s41598-025-93588-z","workflowStages":[]},"version":"v1","identity":"rs-4675940","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4675940","identity":"rs-4675940","version":["v1"]},"buildId":"WrCJVZZCHTDjtuVLN7oU0","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}