To Beep or Not to Beep? Comparing Abstract versus Language-Based Multimodal Driver Displays

摘要

Multimodal displays are increasingly being utilized as driver warnings. Abstract warnings, without any semantic association to the signified event, and language-based warnings are examples of such displays. This paper presents a first comparison between these two types, across all combinations of audio, visual and tactile modalities. Speech, text and Speech Tactons (a novel form of tactile warnings synchronous to speech) were compared to abstract pulses in two experiments. Results showed that recognition times of warning urgency during a non-critical driving situation were shorter for abstract warnings, highly urgent warnings and warnings including visual feedback. Response times during a critical situation were shorter for warnings including audio. We therefore suggest abstract visual feedback when informing drivers during a non-critical situation and audio in a highly critical one. Language-based warnings during a critical situation performed equally well as abstract ones, so they are suggested as less annoying vehicle alerts.