Exploring the Influence of Driving Context on Lateral Driving Style Preferences: A Simulator-Based Study

TL;DR

This work evaluates lateral driving style preferences for autonomous vehicles on rural roads, considering different weather and traffic situations and could not confirm the hypothesis that people prefer to be driven by mimicking their own driving behavior.

Abstract

Technological advancements focus on developing comfortable and acceptable driving characteristics in autonomous vehicles. Present driving functions predominantly possess predefined parameters, and there is no universally accepted driving style for autonomous vehicles. While driving may be technically safe and the likelihood of road accidents is reduced, passengers may still feel insecure due to a mismatch in driving styles between the human and the autonomous system. Incorporating driving style preferences into automated vehicles enhances acceptance, reduces uncertainty, and poses the opportunity to expedite their adoption. Despite the increased research focus on driving styles, there remains a need for comprehensive studies investigating how variations in the driving context impact the assessment of automated driving functions. Therefore, this work evaluates lateral driving style preferences for autonomous vehicles on rural roads, considering different weather and traffic situations. A controlled study was conducted with a variety of German participants utilizing a high-fidelity driving simulator. The subjects experienced four different driving styles, including mimicking of their own driving behavior under two weather conditions. A notable preference for a more passive driving style became evident based on statistical analyses of participants' responses during and after the drives. This study could not confirm the hypothesis that subjects prefer to be driven by mimicking their own driving behavior. Furthermore, the study illustrated that weather conditions and oncoming traffic substantially influence the perceived comfort during autonomous rides. The gathered dataset is openly accessible at https://www.kaggle.com/datasets/jhaselberger/idcld-subject-study-on-driving-style-preferences.

Figures (5)

• Figure 1: a) The Advanced Vehicle Driving Simulator features a six-degree-of-freedom motion platform, a series car cabin, and seven laser projectors. b) Adverse weather conditions in conjunction with oncoming traffic from the driver's perspective. A tablet positioned at the infotainment location is utilized to assess on-drive comfort.
• Figure 2: Overview of the cascaded longitudinal and lateral vehicle control function. Based on the curvature $K$, velocity $v$, and the longitudinal distance to the closest oncoming traffic object $d_{Traffic}$, the target distance to center lane value $d_{CL}$ is calculated. This intermediate result is utilized by the Path Following Model together with the target velocity $v_{Target}$ to calculate the steering wheel angle $\delta$ and the longitudinal acceleration $a_x$. The red-highlighted parameter vectors denote the variables that are adjusted to realize the distinct driving styles.
• Figure 3: Comparison of the different cornering behaviors of the three driving styles. In a) the velocity $v$, lateral acceleration $a_y$, and the distance to the lane-center $d_{CL}$ are shown for a left curve without oncoming traffic. The sportive driving style shows the highest velocity, lateral acceleration, and curve-cutting values based on the parameterization. Constrained by the maximal acceleration values, the path-following model cannot hold the vehicle perfectly in the lane-center, which reduces the difference between the passive and rail driving styles in terms of curve cutting. In b), the reaction to an oncoming truck on a straight road segment is illustrated. The upper part of the figure shows the detected distance between the ego and the target vehicle. When the vehicle is detected, the passive and sportive driving style reduces their distance to the lane-center. In this case, the dashed lines represent the target values of the adaptive curve cutting module, and the solid lines represent the actual measured values achieved by the path after considering acceleration limits.
• Figure 4: Scaled mean ratings on the after-drive inventories tia and arca for the two weather conditions based on \ref{['tab:inventoryResponses']}. For a unified presentation, the items "Vehicle Control", "Stress Level", "Behavior Predictability", "Ride Comfort", and "Satisfaction" of the ARCA questionnaire are scaled using a maximum score of ten. The remaining items of the tia inventory are scaled using a maximum score of five. The outermost circle represents the optimal achievable fulfillment of all criteria; inversed items are denoted with $^*$.
• Figure 5: Isolated evaluation of weather, traffic, and curve type effects on the on-drive subjective relaxation levels, numerical results can be found in \ref{['tab:relaxResponsesStats']} in the Appendix.