Abstract
A key factor to optimal acceptance and comfort of automated vehicle featuresis the driving style. Mismatches between the automated and the driver preferreddriving styles can make users take over more frequently or even disable theautomation features. This work proposes identification of user driving stylepreference with multimodal signals, so the vehicle could match user preferencein a continuous and automatic way. We conducted a driving simulator study with36 participants and collected extensive multimodal data including behavioral,physiological, and situational data. This includes eye gaze, steering gripforce, driving maneuvers, brake and throttle pedal inputs as well as footdistance from pedals, pupil diameter, galvanic skin response, heart rate, andsituational drive context. Then, we built machine learning models to identifypreferred driving styles, and confirmed that all modalities are important forthe identification of user preference. This work paves the road for implicitadaptive driving styles on automated vehicles.