An Autonomous Driving Framework With Self-Configurable Vehicle Clusters

In this paper we propose a new theoretical - yet applicable - framework to model the vehicular traffic. We model a vehicle as an automaton that has its own propulsion and can see the state of other automata in a constant size neighborhood. The rules guiding the change of states of each vehicle comprises of the common traffic rules. By observing the dynamics of such automata, we can devise optimal rules that may relieve the traffic congestion, and increase road safety. Last but not least, this model is an algorithmic framework to devise novel algorithms that use vehicular networks and communication between cars and the infrastructure.