Usability Principles for Improving Voice Assistant of IVIS

Given the proliferation of artificial intelligence-based voice assistant in vehicles, multimodal interaction can be performed in multi-task situation which can affect the primary task quality. Therefore, the usability of multimodal interface design in In-Vehicle Infotainment Systems (IVISs) need to be considered to improve the quality of the primary task and lower the cognitive load. The objective of this study is to develop usability principles for multimodal interface (voice and graphical) design of artificial intelligence-based voice assistant in multitasking environments and provide design guidelines in IVISs. We first developed multimodal interface usability evaluation principles and a checklist. To validate the usability principles and identify undefined usability problems, we conducted heuristic evaluation AI based voice assistant in simulated multitasking environment. We then performed usability evaluation of the newly prototyped assistant that was developed based on the guideline. The result clearly demonstrated that renewed voice assistant multimodal interface that was conducted based on the guidelines had positive impact on usability evaluation compared to conventional voice assistant multimodal interface. We expect that our principles will provide effective guidelines for the AI based voice assistant in the multitasking environment.