Treffer: Clustering algorithms to further enhance predictable situational data in vehicular ad-hoc networks

The modern world is constantly in a state of technological revolution. Everyday some new technological idea, invention, or threat emerges. With modern computer software and hardware advancements, we have the emergence of more internet-enabled devices - or, Internet of Things (IoT) devices. We can now create large networks with any device to gather real-time information about an environment. In conjunction, modern car companies across the board have a push from public demand for a fully-autonomous car. In order to accomplish autonomy safely and effectively, Vehicular Ad-Hoc Networks (VANETs) must be established for a local group of cars and their environment to ensure all correct and relevant information is communicated throughout the network. The data collected in a VANET can be passed to machine learning models in order to predict possible conditions and detect anomalies. This thesis explores different ways of clustering local groups of vehicles along with machine learning algorithms to predict where vehicles are likely to be and detect false or impossible information.