Intelligent driving diagnosis based on a fuzzy logic approach in a real environment implementation

Abstract

This paper considers the problem of diagnosing people's driving skills under real driving conditions using GPS data and video records. For this real environment implementation, a brand new intelligent driving diagnosis system based on fuzzy logic was developed. This system seeks to propose an abstraction of expert driving criteria for driving assessment. The analysis takes into account GPS signals such as: position, velocity, accelerations and vehicle yaw angle; because of its relation with drivers' maneuvers. In that sense, this work presents in the first place, the proposed scheme for the intelligent driving diagnosis agent in terms of its own characteristics properties, which explain important considerations about how an intelligent agent must be conceived. Secondly, it attempts to explain the scheme for the implementation of the intelligent driving diagnosis agent based on its fuzzy logic algorithm, which takes into account the analysis of real-time telemetry signals and proposed set of driving diagnosis rules for the intelligent driving diagnosis, based on a quantitative abstraction of some traffic laws and some secure driving techniques. Experimental testing has been performed in driving conditions. All tested drivers performed the driving task on real streets. The testing results show that our intelligent driving diagnosis system allows quantitative qualifications of driving performance with a high degree of reliability.