Machine Learning and Big Data Processing in a Human-Vehicle Interaction System

Machine Learning and Big Data processing are the key points in the development of an Advanced Driving Assistance System (ADAS) and in general of an Autonomous or Semi-Autonomous vehicle. With the rising of the Internet Of Things (IoT) every object in a road environment will be connected and will interact with all the other components in the scenario. In this scenario, intelligent vehicles will be the linking point for human beings with the general system. Taking all these data coming from the environment, the vehicle will acquire knowledge by means of Machine Learning techniques, in order to improve safety, that is the final goal of the ADAS. In this project all these considerations were taken into account: the creation of a Big Data simulated environment, the processing of these data by means of the Ontologies creation and finally the application of 2 ML techniques thought for obstacle identification, classification and avoidance. This thesis shows all the results achieved, highlighting the previous system problems with the corresponding solutions. As a research thesis, the contribution is related to the ADAS and ML anatomy comparison and analysis, Ontology creation for different scenarios and ML application for the obstacle context. This thesis was a contribution to the artificial intelligence sub-system of a complete ADAS. Final goal of the project is the real implementation of the complete Autonomous Driving system, where the AI sub-system will play a key-role together with a communication intra-vehicle one and a general networking based sub-system.