Effects of Multimodal Warning Types on Driver's Task Performance, Physiological Data and User Experience

Previous studies have compared different multimodal warning types, however, few researchers studied the effects of different multimodal warning types on drivers' task performance, physiological data, and user experience. In our research, we designed a simulated driving experiment to investigate these effects. In the experiment, a small projector, two Bluetooth speakers and two vibration generators were used as signal generators to simulate 6 multimodal warning types, and Mean Deviation, braking reaction time (BRT), normalized GSR, normalized HR, trust, annoyance, satisfaction were used as dependent variables to reflect effects of multimodal warning types on driver's task performance, physiological data and user experience. The main conclusions are drawn in this paper as following: (1) In terms of task performance, there was no significant difference in the effect of different multimodal warning types on driving tasks, but compared with unimodal warning types, both bimodal warning types and trimodal warning type can reduce BRT. (2) In terms of physiological data, normalized GSR and normalized HR were increased by increasing numbers of modalities of warning types. (3) In terms of user experience, trust and satisfaction of multimodal warning types were significantly higher than the other two unimodal warning types. Moreover, the annoyance of warning types which included tactile modality were significantly higher than the warning types which did not included. We speculated that adding tactile signals may increase annoyance of participants, nevertheless, tactile signals still has great potential to increase the trust and satisfaction of multimodal warning types.