Optimizing Hidden Markov Model with Baum-Welch Algorithm for Vehicle Driver's Intention Recognition

Abstract

In this paper, a method for vehicle driver’s intention recognition is investigated through an optimized hidden Markov model (HMM). As uncertainties exist between modeling and application of intention recognition with HMM, the Baum-Welch (B-W) algorithm is introduced to optimize and update the parameters of the established HMM. The major steps within the optimization process, such as acquisitions of forward probabilities, backward probabilities, maximum expected probabilities ratio, elements updates of matrixes π, A and B, and setups of recursion terminated condition, are fully described in the paper. An elaborate example presented at the end of the paper shows the credibility and effectiveness of the proposed algorithm.