Dynamic Network Loading Using Cell Transition Model

Abstract

Static traffic assignment models cannot capture traffic dynamics. Time varying flows, queuing, are not accounted in static models hence traffic dynamics. Hence to replicate the real time scenario Dynamic traffic assignment models are required. Loading the network dynamically gives realistic picture of traffic dynamics which helps for an effective management of traffic congestion hence real picture of incidents all through the network. In this project Cell transmission model (CTM) which is macroscopic and dynamic in nature is used for loading the network dynamically over space and time to capture the traffic dynamics which cannot be captured by using static models to load the network. To investigate how results are affected when a natural incident occurs like sudden increase in demand, vehicle break downs on the link is modelled. Considering each link into number of cells depending upon the free flow speed of typical vehicle in a tick of clock accommodating slower vehicle in the model; the network is loaded dynamically on a three junction road study network consisting ten links to model how the traffic responds to an incident happening in next cell. Simulation for each of the ten links is done and modelled outflow is done and compared with observed outflow. Modelling is done and the network is tested for various incidents depicting the real time scenario, like when traffic enters a cell beyond the capacity flow rate how bottle necks are formed, hence performance of link can be tested. Progression of cell occupancies for different step lengths is modelled and state of a cell is also replicated which is useful to know the real time scenario at a specified cell of a road link at a particular instance of time. The links are modelled and analysed for increased inflow demand, vehicle break downs, capacity drops hence fluctuations in traffic which may be used as real time information for a commuter to make a route choice with help of a user interface. Also for traffic engineer helps in effective planning and management of traffic congestion. Finally we conclude and outline that modelling traffic dynamically effectively tests even the performance of the road network replicating realism which would not be possible with static dynamics, moreover cell transmission model is easier to implement for its simple calculations of linear relationships. Testing and analyzing road network is effectively done.