A fuzzy logic model to estimate safe driving behavior based on traffic violation

Reports from international organizations estimate that death by traffic accidents will increase over the next two decades. One of the main causes of traffic accidents is the traffic violation to road safety norms. Thus, we propose a method to automatically estimate driver's traffic violation from the vehicle's data and traffic signs information. Fifteen drivers traveled during 25 minutes approximately in a selected route, with 24 recognized traffic signs. From these data, human expertise and the Organic Integral Criminal Code (COIP) information, we create a fuzzy logic model to estimate the traffic violation of drivers. To validate our model, we count all contraventions associated with the penalty it would cause to the driver license's points, in the same way, a driver would be issued a traffic ticket. Results show a high correlation between the estimation of the traffic violation and contravention points. This methodology serves as a baseline for real-time estimation of traffic violations. Further work will validate this methodology with more drivers to generalize these results.