Development of Machine Learning techniques for the creation of new fault prevention algorithms.

The aim of this dissertation is to bring together the work carried out on the development of machine learning techniques for the creation of new fault prevention algorithms within Iveco heavy range vehicles. In particular, the Additional Water Heater is considered: the goal is to foresee its potential lock twelve hours in advance. The dissertation is structured in seven chapters and relative paragraphs, which order is in according to the actual path followed during the work development. In “Chapter 1”, an overview of Iveco is offered both from a historical and organizational point of view. In “Chapter 2”, all the features related to the new generation vehicles are described, stressing the role of telematics and connectivity made possible by a new electrical system. In “Chapter 3” the Additional Water Heater is described: an overview about its components is given and the working logics are analyzed. In “Chapter 4” the basics of machine learning theory are introduced along: several issues underlying the field are addressed, along with the techniques involved. In “Chapter 5” machine learning techniques are applied to an actual case study: the Additional Water Heater. Exploratory data analysis is carried out on data coming from telematics in order to get insights, and then a model has been built in order to forecast the component lock twelve hours in advance. In “Chapter 6” the achieved results are presented along with some possible improvements. Finally, in “Chapter 7”, some Python code lines are reported: only the most important customized functions or ad-hoc methodologies are considered in order to show how the obtained results are actually achieved. The presented work is entirely attributable to the author of the dissertation: the discussed analysis and the entire algorithm are the result of a six-months internship during which the author played a crucial role in the development of the project. The author was constantly assisted by the Iveco Team to integrate the obtained results into the ecosystem and to undertake technical analysis directly on the vehicles.