A semantic approach to driving behavior analysis

The aim of this work is to propose a semantic approach to driving behavior analysis. The analyzed vehicle motion parameters are accelerations and positions. Both are sampled using an embedded low-cost lightweight architecture. The authors start from a linguistic description of the problem ontology demonstrating that the selected parameters are sufficient to describe various aspects of the driving behavior. The system is modeled in terms of fuzzy rules and a Fuzzy Inference System is used to classify the driving behavior. The experiments carried out show that the proposed system is able to discriminate among various driving behavior.