Intuitive interaction has long been seen as a highly user-friendly method. There are attempts to implement intuitive interfaces in vehicles in both research and industrial, such as voice commands. However, there is a lack of exploration in the in-vehicle multi-modal intuitive interaction, especially under a dynamic driving scenario. In this research, we conducted a design workshop (N=6) to understand user's needs and designers' considerations on the in-vehicle multi-modal intuitive interface, based on which we implemented our design on both a simulator and a real autonomous vehicle using Wizard-of-Oz. We conducted a user experiment (N=12) on the simulator to explore determinants of users' acceptance, experience, and behavior. We figured that acceptance was significantly influenced by six determinants. Drivers' behavior has an obvious pattern of change. Drivers have been proven to have less workload but distractions were also reported. Our findings offered empirical evidence which could give insights into future vehicle design.
