Effect of Environmental Factors and Individual Differences on Subjective Evaluation of Human-like and Conventional Automated Vehicle Controllers

Abstract

Achieving optimal performance in human-machine systems, such as highly automated vehicles, relies, in part, on individuals’ acceptance and use of the system, which is in turn affected by their enjoyment of engaging with, or experiencing, the system. This driving simulator study investigated individuals’ real-time subjective evaluation of four different Automated Vehicle (AV) driving styles, in different environmental contexts. Twenty-four participants were recruited to manually drive a contextually rich simulator environment, and to experience human-like and non-human-like AV driving styles, as well as the automated replay of their own manual drive. Their subjective real-time feedback towards these driving styles was analyzed. Our results showed that participants gave higher positive feedback towards the replay of their own drive, compared to the other three controllers. This difference was statistically significant, when compared to the high-speed controller (named as Fast), particularly for sharp curves. With respect to the replay of their own drive, participants gave higher negative feedback when navigating an Urban environment, compared to Rural settings. Moreover, changes in roadside furniture affected individuals’ feedback, and this effect was more prominent when the vehicle was driving closer to the edge of the road. Based on our results, we conclude that individuals’ perception of different AV driving styles changes based on different environmental conditions, including, but not limited to, road geometry and roadside furniture. These findings suggest that humans prefer a slower human-like driving style for AV controllers that adapts its speed and lateral offset to roadside objects and furniture. Investigating individual differences in AV driving style preference showed that low Sensation Seeking individuals preferred the slower human-like controller more than the faster human-like controller. Consideration of this human-centered feedback is important for the design of future AV controllers, to enhance individuals’ ride experience, and potentially improve acceptance and use of these vehicles.

Drivers’ Evaluation of Different Automated Driving Styles: Is It Both Comfortable and Natural?

“…This was especially the case for the rural sections, which contained roads of tighter curvature, and higher speed. These results are in line with the work of Hajiseyedjavadi et al (2021), who found that their model-based human-like AV-driving controllers were assessed as less pleasant when negotiating narrower curves. These authors also found that a more rigid controller, which always followed the centre of the lane, received better evaluations.…”

“…The other may be because of the lower operating speed of the AV used in an urban road, by Oliveira et al, compared to the higher travelling speeds of our vehicle, travelling in rural road sections. This contrast in results illustrates the importance of considering the scenarios used to evaluate AV-driving styles in such studies, since they vary across different road environments, based on both geometry and posted speed limit, which clearly influences any subjective assessment and evaluation (Hajiseyedjavadi et al, 2021). Further work on the influence of different scenarios on subjective appraisals of human- vs machine-like AV-driving styles, should clarify this.…”

“…Taking motion planning as an example, the generated driving behaviour of AVs could be robotic, with algorithm-optimised trajectories, based solely on sensory information provided by lasers, radars and cameras, to adapt to the environment (e.g., Urmson et al, 2008). Alternatively, these may mimic a human driver’s average behavioural patterns, by training models, based on real human driving data (e.g., Hajiseyedjavadi et al, 2021; Rehder et al, 2017; Wei et al, 2019). Personalisation of driving styles can also be achieved by using users’ own driving style in the model development loop (e.g., Menner et al, 2019).…”

“…Users’ perception of an AV’s driving style is known to be influenced by both objective and subjective factors. For example, road furniture and geometry are known to influence ratings of safety and comfort (Hajiseyedjavadi et al, 2021) and physiological response (Beggiato et al, 2019; Radhakrishnan et al, 2020), while a number of studies have shown a correlation between personality traits such as Sensation Seeking (Arnett, 1994) and preferred driving style. For example, in manual driving; drivers with high sensation seeking scores are found to drive in a riskier and more aggressive manner and at higher speeds, while low sensation seekers have a tendency to drive more slowly (Louw et al, 2019; Taubman-Ben-Ari et al, 2004; Zuckerman & Neeb, 1980).…”