Predictive Fuzzy Markov Decision Strategy for Autonomous Driving in Highways

This paper presents a model for lane changing in highway driving. It consists of two main parts; threat assessment by assessing the interaction between traffic participants captured in terms of fuzzy logic and a decision making approach based on the concept of Markov Decision Process (MDP). The combined system forms a predictive Fuzzy Markov Decision Process (FMDP). A model predictive control (MPC) scheme for trajectory generation/control complements this decision process. Multiple driving scenarios are evaluated in a visual simulation environment to verify the potential use of the proposed method and reported in the paper.