Real-time implementation of estimation method for driver´s intention on a driving simulator

In recent years, in order to reduce traffic accidents, several kind of support systems for driving safety have been put to practical use. Almost the support systems avoid the dangerous situation based on the recognition results of outside environment of the car. On the other hand, more than 90 % of traffic accidents are caused by human errors. In the research field of human - machine interaction, to estimate and recognize of human psychological of emotional state is needed as one of important fuction for machine side. It is thought that these system are effective as fundamental instrumentation to realize the driving support system, and to avoid the human errors and overconfidence for these support system. The authors have been researched about driver modeling based on the Hidden Markov Model, and its application for the estimation of driver´s intention. In the previous paper, decision methodology of reasonable true value of driver´s intention in the experiment had been proposed. In this paper, implementation of real-time estimation for driver´s intention on the driving simulator system with higher realistic has been introduced, and their validity have been discussed with experimental results.