Driver Acceptance of Adaptive Cruise Control and Active Lane Keeping in Five Production Vehicles

Transcription

1 Driver Acceptance of Adaptive Cruise Control and Active Lane Keeping in Five Production Vehicles March 2017 Ian J. Reagan, David G. Kidd, and Jessica B. Cicchino Insurance Institute for Highway Safety

2 Driver Acceptance of Adaptive Cruise Control and Active Lane Keeping in Five Production Vehicles Ian J. Reagan, David G. Kidd, and Jessica B. Cicchino Insurance Institute for Highway Safety Little is known about how consumers interact with driving automation technology that controls steering, speed, or headway in production vehicles. Forty-eight Insurance Institute for Highway Safety employees used a Honda Civic, Infiniti QX60, Toyota Prius, or Audi A4 or Q7 as a personal vehicle for up to several weeks and completed surveys about their experiences. Agreement about whether adaptive cruise control (ACC) or active lane keeping (ALK) improved driving experience varied significantly among vehicles. The Q7 s ACC improved the driving experience significantly more than its ALK. The Civic s ALK improved the driving experience more than the Q7 s system, but this effect only approached significance. Drivers were most comfortable using systems on free-flowing interstates and least comfortable using ACC in stopand-go traffic and ALK on curvy roads. The findings show a range of qualitative differences in driving automation technologies and that use of current technologies likely is limited to low-demand conditions. Introduction Driver assistance and collision warning technologies have the potential to prevent a significant proportion of crashes (Jermakian, 2011). Some of these systems, such as front crash prevention, are reducing insurance claim frequency and police-reported crash rates (Cicchino, 2017; Highway Loss Data Institute ((HLDI), 2015; HLDI, 2016a), but mixed results have been observed for others, such as lane departure warning (HLDI, 2016b; HLDI 2016c; Sternlund, Strandroth, Rizzi, Lie, & Tingvall, 2017). Some technologies may not be living up to their potential in part because drivers are not using them (e.g., Reagan & McCartt, 2016). Driver trust and annoyance vary among driver assistance technologies (Eichelberger & McCartt, 2014; Kidd, Cicchino, Reagan, & Kerfoot, 2016), and these differences in the way drivers interact with and experience these technologies may influence their use. Driving automation technologies build on current driver assistance systems. Unlike crash avoidance technologies that intervene in safety critical situations, driving automation technology adapts to developing situations and may prevent critical situations from happening altogether. Active lane keeping (ALK) and adaptive cruise control (ACC) are two forms of driving automation technology available in production vehicles. ACC provides longitudinal vehicle control to maintain speed and headway. ACC may convey benefits similar to or in parallel with front crash prevention (Kessler et al. 2012). Field operational tests and simulator studies show increased following distance, gentler braking, or reduced speeds when ACC is engaged relative to when it is not (e.g., Kessler et al., 2012; Vollrath, Schleicher, & Gelau, 2011). ALK provides sustained steering input to keep the vehicle within its travel lane. Drivers may potentially find the increased functionality of this system more useful than lane departure warning, especially if it minimizes alerts that drivers find annoying (e.g., Eichelberger & McCartt, 2014). The potential safety benefits of driving automation technology depend on how frequently and where it is used. Drivers may not use technologies like ACC and ALK frequently or may limit use to situations where there is little opportunity for the technologies to benefit safety. For example, drivers do not use ACC all the time and use tends to be highest during free-flowing conditions on higher speed roadways where forward conflicts are less frequent (General Motors Corporation, 2005; Kessler, Etemad, Alessandretti, et al., 2012). Also, drivers may use driving automation technology even less if they think it lacks value or is unpredictable (Beggiato & Krems, 2013; Jamson, Merat, Carsten, & Lai, 2013). The purpose of this study was to learn more about driver perceptions of value and utility of ACC and ALK following on-road use. Specifically, drivers were asked if these technologies improved their driving experience and about situations in which they felt comfortable using them. Method Participants Participants were Insurance Institute for Highway Safety (IIHS) employees in Arlington or Ruckersville, Virginia. Arlington is a densely populated area outside Washington, DC, and Ruckersville is a rural area about 20 miles north of Charlottesville, VA. Of the 108 employees who received an invitation, 47 agreed to voluntarily participate (Table 1). No incentives were offered to participate. Table 1. Sample demographics overall and by location Location Age Gender Mean SD Min Max M (n) F (n) Arlington Ruckersville Overall Vehicles Participants drove up to five different vehicles. They were a 2017 Audi A4, 2017 Audi Q7, 2016 Honda Civic, 2016 Infiniti QX60, and 2016 Toyota Prius. All vehicles except the A4 were used in a previous study involving IIHS employees, so about 1/3 of participants in this study had prior experience with a vehicle they drove. Every vehicle was equipped with multiple driver assistance systems, but only the driving automation technologies are discussed. Each vehicle was equipped with an ACC system capable of bringing the vehicle

3 to a complete stop. Each ACC system had controls to set and adjust the vehicle s speed and headway to a vehicle ahead; however, the layout and location of these controls varied. A unique feature of Audi s ACC system was a speed limit recognition function that adjusted the set speed using camerabased technology. The feature slowed the vehicle when ACC was engaged and traveling faster than speed indicated on speed limit signs. Three (Audi A4, Audi Q7 and Honda) of the five vehicles were equipped with ALK. The Audi ALK systems operated at speeds over 40 mph (A4 and Q7), and the Honda system became available at 45 mph (Civic). Both Audis were equipped with a congestion assist feature that provided active lane keeping and car following at slower speeds. In addition to ALK, both Audis and the Honda had a lane departure prevention system that provided a haptic warning (vibrating steering wheel) and corrective steering when the vehicle departed the lane. Procedure Researchers contacted participants to schedule days to use a vehicle. Participants signed vehicle use agreements where they agreed to use the vehicle instead of their personal vehicle, follow traffic laws, avoid cellphone use, and drive with the technologies engaged as often as possible. Participants completed a familiarization process prior to receiving the vehicle. First, a researcher showed participants where various controls (ignition, seat and mirror adjustment controls, windshield wipers) were and how to use them. Then the researcher summarized each crash avoidance and driver assistance technology equipped to the vehicle. The summary included descriptions of each systems purpose, operational boundaries, settings, and instruction about use (e.g., activate, deactivate, change settings). All drivers received vehicles with systems activated except for those that required activation when the vehicle was moving. The researcher also made sure that system settings were set to a middle sensitivity setting or the earliest setting if only two were available during this overview. Next, participants used ACC and ALK during a 30- minute researcher-supervised drive on the interstate. The researcher instructed the participant to engage ACC, modify the set speed, and adjust the headway setting. The researcher encouraged each driver to allow the ACC system to accelerate and decelerate the vehicle in response to prevailing traffic conditions. Participants were instructed to depart the travel lane in a controlled manner when immediate traffic was absent and a wide, paved shoulder was present. This maneuver was intended to demonstrate the differences between active lane keeping, lane departure warning, and lane departure prevention. Status and indicator displays were highlighted during use. Drivers completed an online daily-use survey each day that they used the vehicle. This survey took about 5-10 minutes to complete and asked participants about their driving exposure (e.g., miles driven; percentage of driving time on higher speed roads, in free-flowing traffic, and during precipitation). Drivers completed a post-use survey after returning the vehicle. The survey collected information about participants experiences and opinions about the various vehicle technologies by asking participants to indicate their level of agreement or disagreement with various statements. Participants indicated their level of agreement using a 5-point Likert scale that ranged from 1 (strongly disagree) to 5 (strongly agree). The following three statements about ACC and ALK were intended to capture drivers general sentiments of each system: Overall, I felt this [adaptive cruise control/active lane keeping] system improved my driving experience. I want this [adaptive cruise control/active lane keeping] system on my next car. This [adaptive cruise control/active lane keeping] system worked as I expected. Participants also responded to additional statements about their comfort with using ACC or ALK in different roadway scenarios. These statements were based on feedback received in a previous study (Kidd et al., 2016) about situations where ACC and ALK systems were used, had difficulty, or behaved unexpectedly. Drivers rated their level of agreement with whether they felt comfortable using ACC in free-flowing traffic on interstates, free-flowing traffic on major arterial roads with signalized intersections, roads with moderate hills, heavy stop-and-go traffic, and low-speed local roads. Drivers rated their level of agreement with whether they would feel comfortable using ALK on winding, curvy roads; roads with moderate hills; on interstates with gentle to moderate curves; and in free-flowing traffic on interstates. The post-use survey took about 30 minutes to complete. Data for this study was collected from August 2016 to January Analysis Plan Responses to the three statements about drivers general sentiments toward ACC and ALK were analyzed using mixed-effects regression. The level of agreement with each statement was modeled separately with the fixed-effects of vehicle (Audi A4, Audi Q7, Honda, Infiniti, Toyota), technology (ACC, ALK), and the interaction between vehicle and technology. Separate mixed-effects regression models were constructed to examine differences in the level of agreement that drivers were comfortable using ACC and ALK in different driving situations. These two models included the fixed effects of situation, vehicle, and the interaction between situation and vehicle. Driver was included as a random effect in every model to account for within-subject variance from repeated observations. Type 3 tests were performed to determine if the set of variables that made up each fixed effect was statistically significant at the 0.05 level. Least Square Means (LSM) estimates and 95% confidence intervals were computed to examine pairwise differences among the levels in each fixed effect and are reported after each statistical test. Pairwise comparisons were adjusted using the Tukey-Kramer method. Analyses were conducted with SAS 9.4 using PROC MIXED. Results

4 Driving Exposure Participants used the study vehicles 80 times. Twenty-four used only one vehicle; 15 used two; seven used three; and one driver used all five vehicles. Vehicle use ranged from 1-15 days and was six days on average. Drivers submitted 423 daily-use reports and reported driving 31,331 total miles. Table 1 shows the number of drivers, daily reports, and total mileage reported for each vehicle. Table 1. Driving exposure with study vehicles. Vehicle Drivers (n) Daily reports (n) Reported miles driven 2017 Audi A , Audi Q , Honda Civic , Infiniti QX , Toyota Prius ,107 General sentiments about ACC and ALK The average level of agreement with statements about driver sentiments toward ACC and ALK are shown in tables 2 and 3. Across the five vehicles, drivers agreed that ACC improved their driving experience (M=3.90), wanted ACC on their next vehicle (M=3.99) and worked as expected (M=4.00). When asked about ALK, drivers average level of agreement with these statements was lower, ranging from 3.33 to Given the collinearity between responses to these three statements, only the statement Overall, I felt [adaptive cruise control/active lane keeping] improved my driving experience was analyzed further. Sixty-nine percent of participants agreed or strongly agreed that adaptive cruise control improved their driving experience, and only 51% said the same about active lane keeping. The average level of agreement with this statement for each vehicle with ACC and ALK is shown in Figure 1. A mixed-effects regression did not indicate a significant main effect of vehicle (F(4, 115)=0.74, p=0.57) or main effect of technology (F(1, 115)=2.5, p=0.12), but there was a significant interaction between vehicle and technology (F(2, 115)=6.4, p<0.01). Post-hoc tests indicated that drivers expressed greater levels of agreement that the Audi Q7 s ACC system improved the driving experience than its ALK system (Figure 1) (p=0.02). Drivers also indicated greater levels of agreement that the Honda ALK system improved their driving experience compared with the Audi Q7 s system; however, this difference only approached significance (p=0.09). Table 2. Mean (SD) level of agreement (1=strongly disagree, 5=strongly agree) with general sentiments about ACC overall and by vehicle. Vehicle Improved my driving experience Want on my next vehicle Worked as I expected Audi A (0.83) 4.00 (1.04) 3.64 (0.93) Audi Q (0.70) 4.13 (0.83) 4.20 (0.68) Honda Civic 3.50 (1.16) 3.57 (1.16) 3.86 (1.35) Infiniti QX (0.71) 4.09 (0.75) 4.09 (0.97) Toyota Prius 4.13 (0.83) 4.07 (0.88) 4.13 (0.74) Overall 3.90 (0.85) 3.99 (0.92) 4.00 (0.95) Table 3. Mean (SD) level of agreement (1=strongly disagree, 5=strongly agree) with statements about ALK overall and by vehicle. Vehicle Improved my driving experience Want on my next vehicle Worked as I expected Audi A (0.63) 3.50 (0.85) 3.64 (0.84) Audi Q (1.07) 2.73 (1.16) 3.27 (1.22) Honda Civic 4.00 (1.18) 3.79 (1.31) 3.93 (1.27) Overall 3.53 (1.05) 3.33 (1.19) 3.60 (1.14) Driver responses to these three statements about ACC and ALK were similar, suggesting they provided comparable information. Bivariate mixed-effect regressions were constructed to assess the strength of association between these statements separately for ACC and ALK, while accounting for multiple observations from drivers. The proportion of variance that each statement accounted for in another statement (i.e., R 2 ) was estimated by computing the percentage reduction in the total residual variance observed in each bivariate model from a null model with only the intercept. The square root of the R 2 values was used to estimate the correlation between pairs of statements. The correlations between pairs of statements were all above 0.50, indicating a large effect size (Cohen, 1988), and they ranged between 0.55 and 0.78 for ACC and 0.71 and 0.76 for ALK. The coefficients were nearly identical to Pearson product-moment correlation coefficients that ignored the structure of the data. Figure 1. Least square means estimate for driver agreement that ACC or ALK improved their driving experience. Error bars indicate 95% confidence intervals. Driver comfort with using ACC and ALK Driver agreement with statements about comfort with using ACC in different driving situations was analyzed next. Eighty-eight percent of drivers agreed or strongly agreed that they would be comfortable using adaptive cruise control in free-flowing traffic on interstates, but only 35% said the same about using ACC on low-speed, local roads. The analysis revealed a significant main effect of situation (F(4, 375)=39.85, p<0.001). On average, drivers had greater levels of agreement about being comfortable using ACC in freeflowing traffic on interstates (LSM=4.39, 95% CI[4.14, 4.64]) compared with arterial roads with intersections (LSM=3.64, 95% CI[3.39, 3.89]), roads with moderate hills (LSM=3.65, 95% CI[3.40, 3.90]), in heavy stop-and-go traffic (LSM=3.00, 95% CI[2.75, 3.25]), and low-speed local roads (LSM=2.93, 95% CI[2.69, 3.18]). Drivers additionally indicated significantly greater agreement in being comfortable using ACC on arterials with intersections and roads with moderate hills compared with low-speed local roads or in heavy stopand-go traffic.

5 There also was a significant main effect for vehicle (F(4, 375)=3.98, p<0.01). On average, drivers indicated greater agreement with being comfortable using the ACC system in the Audi A4 (LSM=3.60, 95% CI[3.31, 3.89]), Audi Q7 (LSM=3.94, 95% CI[3.66, 4.22]), and Infiniti QX60 (LSM=3.54, 95% CI[3.29, 3.78]) compared with Honda s system (LSM=3.15, 95% CI[2.86, 3.44]). Additionally, drivers were significantly more comfortable using ACC in the Audi Q7 than the Toyota (LSM=3.38, 95% CI[3.10, 3.66]) and Infiniti. The interaction between vehicle and situation was not statistically significant (F(16,375)=0.69, p=0.81) Like ACC, the largest proportion of drivers agreed or strongly agreed that they would be comfortable using ALK on interstates with free-flowing traffic (67%). Only 40% of drivers agreed or strongly agreed they were comfortable using ALK on curvy, winding roads. A mixed-effect regression modeling drivers level of agreement with being comfortable using the Audi A4, Audi Q7, and Honda ALK systems in different situations indicated a significant main effect of situation (F(2, 160)=7.51, p<0.001). Drivers indicated greater agreement in being comfortable using ALK in free-flowing traffic on interstates (LSM=3.95, 95% CI[3.60, 4.30]) compared with curvy, winding roads (LSM=3.16, 95% CI[2.82, 3.51]). Drivers also had greater agreement with being comfortable using ALK in free-flowing traffic on interstates compared with roads with moderate hills (LSM=3.49, 95% CI[3.14, 3.83]), and were more comfortable using ALK on interstates with gentle to moderate curves (LSM=3.63, 95% CI[3.28, 3.97]) than winding, curvy roads; however, these differences only approached significance (p=0.06). The main effect for vehicle approached statistical significance (F(2,160)=2.82, p=0.06. On average, drivers were more comfortable using the Honda ALK system (LSM=3.84, 95% CI[3.47, 4.21]) across the four driving situations, compared with the Audi A4 (LSM=3.40, 95% CI[3.03, 3.78]) and Audi Q7 s (LSM=3.43, 95% CI[3.07, 3.79]) systems. The interaction between vehicle and situation was not statistically significant (F(6,160)=0.60, p=0.73) Discussion This study evaluated driver opinions and comfort with using ACC and ALK systems in five production vehicles following actual on-road use. The findings show that acceptance of driving automation technology varied across different implementations of the same technology. Every ACC system had similar capabilities, but drivers experienced them differently. Drivers thought Honda s ACC system improved their driving experience the least and that the Audi Q7 s system improved it the most. Opinions about ALK equipped to three of the vehicles also varied; however, the pattern among vehicle models was opposite of ACC. Drivers thought Honda s ALK system improved the driving experience more than the Audi Q7 and Audi A4 systems. Drivers must use driving automation technologies extensively for the safety benefits of the technologies to be realized. However, similar to past research, drivers were more comfortable using driving automation technologies in some situations than others, which suggests drivers restrict their use of the systems (General Motors Corporation, 2005; Kessler et al., 2012). Drivers were most comfortable using ACC and ALK in free-flowing traffic on interstates and least comfortable using ACC in stop-and-go traffic and on local roads and ALK on curvy, winding roads. Most fatal crashes do not occur on interstates (NHTSA, 2014), so the safety benefits of these technologies will be relatively small if they are only used on these roadways. This extends to the types of crashes that these systems can potentially prevent. For example, among passenger vehicle occupants in 2015 killed in singlevehicle crashes, which are relevant to ALK, 36% occurred while the driver negotiated a curve, and 38% occurred on arterials other than interstates (Insurance Institute for Highway Safety, 2017). Furthermore, drivers may benefit from technological assistance the most during more challenging situations where they expressed the least comfort with using ACC and ALK. The current study suggests that the functionality of both technologies needs to mature before drivers are comfortable enough to trust them in situations where they may offer greater safety benefits. A previous study found that driver trust in different driver assistance systems varied across vehicles (Kidd et al., 2016). Likewise, drivers level of comfort with using ACC and using ALK varied across vehicles. Drivers felt most comfortable using the Audi Q7 s ACC and least comfortable using Honda s system. They felt most comfortable using Honda s ALK system and least comfortable using Audi s. This pattern of results also was observed when drivers reported whether the technology improved their driving experience. Kidd et al. (2016) examined driver experiences with 4 of the 5 vehicle models used in the current study and found that drivers trusted ACC more than ALK, and among ACC systems they trusted Honda s system the least. Driver comfort with using a technology and whether the technology improves their driving experience may reflect trust. Trust mediates use when operators interact with complex systems, and trust increases when a system meets users expectations and is helpful in achieving their goals (Lee & See, 2004). The current findings mirror those from Kidd et al. (2016) and suggest that acceptance and use of driver assistance and driving automation technology will vary not only by technology but also by the way it is implemented. Reagan, Cicchino, Kerfoot, and Weast (2017) found that the percentage of vehicles that had lane departure warning, lane departure prevention, and active lane-keeping systems turned on varied widely across vehicles produced by nine automakers. While the authors identified design features, such as warning modality and level of intervention (warn, steer with lane departure, active lane keeping), that affected the likelihood that lane maintenance systems were turned on, the current findings suggest that other factors (e.g., functionality and performance) may contribute to variation in use. Operators need to have an accurate understanding about the purpose, function, and limitations of a system so that their trust and use of the system is correctly calibrated with its capabilities to prevent disuse and misuse (Lee & See, 2004; Parasuraman & Riley, 1997). Owner s manuals are one source of information about driver assistance systems and driving automation. Owner s manuals for current study vehicles provided inconsistent information about the intended

6 operational design domain for ACC and ALK when equipped. For example, the Infiniti and Toyota manuals discouraged use of the ACC in congested traffic and winding, hilly roads; in contrast, the Audi manuals stated that ACC can improve the driving experience in stop-and-go traffic. The Honda manual does not describe the intended operational design domain for its ACC. These differences suggest that drivers could have very different mental models and expectations for a technology with similar functional capabilities. However, the interaction between situation and vehicle on drivers level of comfort with using ACC and ALK in different situations was not statistically significant and suggests that comfort with using driving automation in different situations did not reflect the cautions or intentions specified by the manufacturer in the owner s manual. If drivers do not know or understand the limitations of driving automation technology, then they may use the technology outside the intended operational driving domain. Drivers are discouraged from using Toyota s and Infiniti s ACC in stop-and-go traffic, but ACC works in these conditions, and there is no design feature to prevent the driver from engaging the system. Drivers may assume they are within an operational driving domain because they can engage driving automation technology, and, as a result, may have an adverse experience that impacts future use or even safety. Designers should consider restricting the use of driving automation technology to its intended operational design domain instead of relying on the driver s discretion to prevent misuse or abuse that can have unintended consequences. Conclusion It is widely cited that driver error accounts for about 94% of crashes (NHTSA, 2016). Highly automated vehicles are being promoted as a possible countermeasure for crashes caused by driver error. In the interim, existing driving automation technology may offer safety benefits (e.g., Kessler at al. 2012). The current findings indicate that not all drivers strongly felt that ACC and ALK improved their driving experience. Further, drivers were most comfortable using the technologies in low demand situations, which suggests limited use and safety benefits. Driver sentiments varied by vehicle model, but reasons for the variation remain to be investigated. Acknowledgements The authors thank Laura Kerfoot for assisting with data collection. This work was supported by the Insurance Institute for Highway Safety. References Beggiato, M. & Krems, J.F. (2013). The evolution of mental model, trust, and acceptance of adaptive cruise control in relation to initial information. Transportation Research Part F, 18, Cicchino, J.B. (2017). Effectiveness of forward collision warning and autonomous emergency braking systems in reducing front-to-rear crash rates. Accident Analysis and Prevention, 99A, Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 nd Ed.). Hillsdale, NJ: Lawrence Erlbaum. Eichelberger, A.H. and McCartt, A.T. (2014). Volvo drivers experiences with advanced crash avoidance and related technologies. Traffic Injury Prevention, 15, General Motors Corporation. (2005). Automotive collision avoidance system field operational test (ACAS FOT) final program report (DOT HS ). Washington, DC: National Highway Traffic Safety Administration. HLDI (2016a) Subaru collision avoidance features. HLDI Bulletin, 33(30), HLDI (2016b). Mercedes-Benz collision avoidance features: a 2016 update. HLDI Bulletin, 33(23), HLDI (2016c). Buick collision avoidance features: an update. HLDI Bulletin 33(18): 1 9. HLDI (2015) Honda Accord collision avoidance features. HLDI Bulletin, 32(33), IIHS (2017). Unpublished analysis. Arlington, VA: Author. Jamson, A.H., Merat, N., Carsten, O.M.J., & Lai, C.H. (2013). Behavioural changes in drivers experiencing highlyautomated vehicle control in varying traffic conditions. Transportation Research Part C, 30, Kessler, C., Etemad, A., Alessandretti, G., Heinig, K., Selpi, Brouwer, R., Cserpinszky, A., Hagleitner, W., & Benmimoun, M. (2012). European large-scale field operational tests on in-vehicle systems deliverable 11.3: Final report. Kidd, D.G., Cicchino, J.B., Reagan, I.J., and Kerfoot, L.B. (2016). Driver trust in five driver assistance technologies following real-world use in four production vehicles. Arlington, VA: Author. Lee, J.D. & See, K.A. (2004). Trust in automation: designing for appropriate reliance. Human Factors, 46(1), National Highway Traffic Safety Administration (2014). Traffic Safety Facts 2014 (DOT HS ). Washington DC: Author. National Highway Traffic Safety Administration (2016). The road ahead: National Highway Traffic Safety Administration strategic plan (DOT HS ). Washington, DC: Author. Parasuraman, R. & Riley, V. (1997). Humans and automation: use, misuse, disuse, abuse. Human Factors, 39(2), Reagan, I.J. and McCartt, A.T. (2016). Observed activation status of lane departure warning and forward collision warning of Honda vehicles at dealership service centers. Traffic Injury Prevention, 17, Reagan, I.J., Cicchino, J.B., Kerfoot, L.B., and Weast, R.A. (2017). Crash avoidance and driver assistance system settings in vehicles observed at retailers service departments. Arlington, VA: Author. Sternlund, S., Strandroth, J., Rizzi, M., Lie, A. and Tingvall, C. (2017). The effectiveness of lane departure warning systems a reduction in real-world passenger car injury crashes. Traffic Injury Prevention, 18, Vollrath, M., Schleicher, S., & Gelau, C. (2011). The influence of cruise control and adaptive cruise control on driving behaviour a driving simulator study. Accident Analysis & Prevention, 43(3),