User-adaptive interface based on mental model and symbol matching

We design a user-adaptive interface that is adapted to a driver considering his/her personal differences. This paper deals with the interactive adaptation interface based on probability model of hidden Markov model and symbol matching for in-vehicle information systems. This interface builds a user's navigation pattern model by navigation data and recommends symbols or clues which are affected by the user model. User's intention is inferred by the interactive behavior between the driver and the interface which is symbol matching. As an application example, we present a car navigation system and experiments were performed to confirm the effectiveness on the reduction of the response time.