Nowadays traffic congestions are one of the major problem of urban traffic systems. Two main causes can be discerned. Firstly the configuration of traffic lights is not aware of the current traffic demand (direction and height of the flow). Secondly there are too many unnecessary changes of lanes.
Making an intelligent and adaptive traffic controller system, which calculates with the actual traffic demand, can be a proper solution to the first mentioned problem.
Moreover forming groups of the vehicles and treating them like a single entity can be the solution to the second problem.
In my Bsc thesis I examine these proposed solutions as a complex, intelligent agent based system. I work out a simple communication protocol between the vehicles themselves and between vehicles and the infrastructure.
I verify its merits with simulations based on traffic situations in Budapest as well as in a fictional American style road-network by a self-made simulation platform, which is based on the open-source Simulation of Urban Mobility traffic simulator.