Towards Proactive Adaptive Vehicle Settings

In recent years, cars are increasingly computerized, where the handling of the vehicle can be changed to accommodate individual needs. One specific feature in current vehicles that can alter the vehicle’s dynamic behavior are driving modes: predetermined vehicle settings that drivers can select by the press of a button. Unfortunately, user studies showed that the option to switch modes is underutilized. Possible explanations include mode confusion: drivers may not know when certain vehicle settings could be used best, or they may simply forget the current mode (or forget to change mode). Besides changing driving modes when the vehicle is stationary, driving modes offer the possibility to switch while driving. In theory, this could mean that during a sportier maneuver, such as curve driving or an overtaking maneuver, the driver benefits from dynamic vehicle settings. However, in practice, it is unlikely that drivers will select their preferred vehicle setting in dynamic driving situations or for short periods. A system that automatically changes the vehicle settings for the driver could potentially solve these issues.

The aim of this dissertation is to provide new quantitative and qualitative insights into the underlying principles to design a system with proactive adaptive vehicle settings: A system that automatically changes the vehicle settings to fit the individual and context-dependent needs of the driver.

The first part of this thesis (Chap 2–4) investigates how people adapt to different road environments (road width and curvatures), task instructions, and car characteristics. This kind of knowledge would help to develop a system that adapts according to what the human driver would want when the location (where they drive), the target (i.e., eco vs. normal vs. sport), or the vehicle changes.

The second part of the thesis (Chap 5–7) investigates how offline changes in vehicle settings (e.g., sound, powertrain settings, steering settings) affect the vehicle's dynamic behavior, driving behavior and driver experience. In this part, these questions are addressed for offline vehicle setting changes: changes that occur between driving trials and not while driving. In this way, transient effects in the data can be removed.

The final part of the thesis (Chap 8–9) combines all the learned principles from the previous chapters and investigates how online changes in vehicle settings affect driving behavior and driver experience.

Finally, the individual contributions of each chapter are integrated towards overarching conclusions, limitations, and future work. In short, five overarching conclusions were drawn:
1. Motivational driving models that use emotions or experiences as a construct are theoretically insightful but impractical; driving behavior could better be predicted by car state or location-specific variables.
2. A large part of the variability in driving behavior can be explained by location; location should be included in the design of an adaptive vehicle setting system.
3. The tested sport mode led to objectively more ‘sporty’ vehicle dynamics.
4. Sport mode settings are clearly perceived but do not cause speeding behavior.
5. Proactive adaptations of vehicle settings can objectively improve acceleration performance, lane-keeping, and steering performance, but are not always accepted by drivers.