This paper presents modulus designs for developing the "smart" in-vehicle route guidance systems and the associated algorithms for the smart features. The purpose of these software-oriented algorithms is to enhance the service functions for better guidance. Most on-market in-vehicle guidance systems provide either shortest route or fastest route towards destination at certain instant of time regardless of either individual's perception/reaction or the interaction between such perception with the associated outcomes to such guidance. The core of the "smart" concept is to integrate the en-route route choice behavior models into the guidance system. The en-route route choice behavior has been defined under a "node-to-node" dynamic framework, which is an extension of the author's researches in day-to-day dynamics. Two possible alternative models are discussed in this paper: one based on techniques of Fuzzy Neural Network (FNN) and the other based on Dynamic Probabilistic Models. A preliminary usability test conducted through a laboratory simulator has shown promising results. (A*)
Abstract