A modeling method for predicting driving behavior concerning with driver's past movements

Abstract

Recently, studies of predicting driving behavior based on behavioral models have been done for constructing Driving Safety Support Systems (DSSS) responding to driver's intention. Although traditional behavioral models predict future behavior by analyzing instantaneous velocity and pedal strokes, past movements should be concerned for accurate prediction since human's behavior is strongly related to past actions. This study proposed a method of modeling driving behavior concerned with certain period of past movements by using AR-HMM (Auto-Regressive Hidden Markov Model) in order to predict stop probability. As results of comparison with a conventional method, our algorithm is effective for predicting driving behavior accurately. ©2008 IEEE.