This project focuses on developing intelligent vehicles. Introducing intelligence in everything around us is the most desirable means of human technology. A system is divided into components called agents where the agents communicate with each other to share their state information and makes decision. Similarly in Traffic control System consists of agents such as vehicles , roads, signals etc.
In this project several traffic models will be build from simple to complex. Each model will model a particular phase of traffic scenario such as car following, lane changing, variable speeds, speed- limit, vehicle composition, signal phase control, traffic light controls etc. Incorporated vehicular communications(V2V, V2I, I2V) decisions based on road traffic density to reduce congestion, and analyzing the performance across each model. The information computed from each model is very helpful to avoid road congestions. The effects priors and later to introducing changes during experimentation could be observed through the graphs. These logics will then be aggregated into complex models such as Open Street Map Models imported from Google-Maps.
A microscopic model is able to simulate traffic in urban areas in real time for use in driving simulators. The vehicles considered in the simulation , namely user-driven vehicle at the center of the simulation model and other vehicles interact with its surroundings. In order to understand vehicle dynamics , a wide range of microscopic models are used like car following model , lane changing model etc. which provides a realistic modeling of driver and vehicle behavior. The primitive models developed here are the basic intersection model of traffic lights. The color changes indicating the permission for one lane to process the instruction one at a time whereas the other lane needs to wait. It is observed that for random traffic along the lanes the waiting time remains regular along both directions. Traffic light alternates between green and red phase. By varying the time of green and red phases, the number of vehicles is controlled and observation are carried out in 6 lanes.
The proposed randomized traffic light controller is capable of communicating with the neighbor junctions and manages phase sequences and phase length adaptively. A real case study of complex traffic junction is simulated having four intersections. Average flow density, average delay time , and linkoverflow of all four intersections are used as performance indices.
The project deals with adding resilience to existing control plane Architecture using fault notification and recovery mechanisms in order to improve scalability of existing networks