Machine Learning Models for Predicting Driver Behavior in Mixed Traffic Scenarios

Abstract

 Machine learning has been widely used in recent years to predict human driver behavior. Various methods and techniques have been applied in different ch severities such as driving behavior prediction, intention prediction, and trajectory prediction, all of which are based on machine learning models. However, due to advancements in intelligent vehicles and artificial intelligence (AI) technology, the interactions in mixed traffic scenarios are becoming more complicated. Although considerable contributions have been made to enable human-like, decision-making behaviors in intelligent vehicles, e.g., improving motion planning and control, there is still a lack of investigation of adopting comprehensive machine learning models to accurately predict and model human driver behaviors. Therefore, developing machine learning models for predicting complex driver behaviors in mixed traffic scenarios with different densities is still a challenging open problem. In this work, we are targeting to build state-of-the-arts machine learning models for predicting and modeling driver behavior in mixed traffic scenarios. And our final target is to develop an intelligent vehicle that can coexist with human drivers harmoniously