Evolution of technology has caused a major increase of urban traffic, which is becoming more and more difficult to handle by using current road network. Besides building new roads, adequate traffic control can increase the throughput of urban links, but this approach needs a suitable framework for modelling and simulating urban traffic.
This Thesis presents such a framework implemented in MATLAB environment. Modelling is based on batch Petri nets, which extends the capabilities of traditional Petri nets and is therefore suitable for modelling such phenomena as common speed changes or congestions, occurring frequently in urban traffic scenarios. The framework also supports testing various traffic control algorithms. The implemented functions have been validated on several test networks.
Models of two common intersection types, namely the signalled cross-intersection and the signalled T-shape intersection have been developed and integrated to the framework. The model of straight road sections connecting the intersections has also been elaborated using a batch place, and therefore being capable of modelling dynamically changing speed and density of vehicle groups.
The Thesis proposes a traffic control algorithm, which is capable of tuning cycle times of traffic lights adaptively, based on current traffic situation. Efficiency of the algorithm is illustrated by simulations carried out on a complex road network.