A Recognition Method for Lane Change Intention Based on Hidden Markov Model

Abstract

In order to reduce drivers' misbehavior and errors, various driving assistance systems (DAS) are designed to provide effective warnings and alarms when dangerous traffic condition occurs. When designing accurate DAS, it is significant to recognize drivers' driving intentions, especially lane change intention, by combining the information of vehicle state and driver's behavior. In this study, a driver's lane change intention recognition method is proposed. The driver's lane change intention is recognized by Hidden Markov Model (HMM), with various inputs of vehicle acceleration, turn signal condition and steering wheel angle. The model is validated by implementation of simulation for verifying the recognition accuracy and efficiency on a driving simulator. The results of experiment indicate that lane change intention can be recognized efficiently by the model established.