Ontology-Based Driving Scene Modeling, Situation Assessment and Decision Making for Autonomous Vehicles

In this paper, an ontology-based driving scene modeling, situation assessment and decision making method for autonomous vehicles in urban environment is proposed. Firstly, an ontology is developed to model the driving scene of urban environment and represent driving knowledge in machine-readable format. Then, a general deterministic situation assessment approach which assesses the safety of eight regions around the autonomous vehicle and legitimacy, reasonableness for changing to adjacent lanes is proposed. Based on the assessment results, the prolog reasoner is employed to generate reasonable behavior decisions. Field tests are conducted and the experimental results show that the proposed method can give effective decisions in different urban scenarios.