A supervised machine learning framework for smart tires

Artificial intelligence techniques are today among the most important challenges for the future due to their ability to analyze big data and identify key characteristics. In transportation, the growing amount of data measured on vehicles and infrastructures constitute a wealth of information useful for energy, safety and environmental sustainability. In this paper, a supervised machine learning methodology applied to smart tires is presented. Starting from the definition of sensors inside the tire, it is shown how a predictive model can be used for the learning phase through the generation of virtual datasets. Following the learning phase, the architecture and the functional scheme of the machine learning algorithm is presented. The methodology is capable to estimate key variables of the tire operation starting from common measurements on the vehicle.