SensorGrid for Real-Time Traffic Management

The combination of sensor networks and grid computing enables the complementary strengths and characteristics of sensor networks and grid computing to be realized on a single integrated platform. As each runs on different platforms and uses different languages, the usage of Web services is the best choice to implement this integration. Stargate, a computer with sensor signal processing capabilities is used in the implementation. This combination called SensorGrid allows powerful applications to be built on it to capitalize on the real-time sensing and the huge computing ability. The application that will be examined in this project is car traffic optimization. Powerful Q-learning algorithms are running in distributed Stargates to vary the duration of green and red light at each traffic junction according to the traffic situation that is being sensed continuously. The benefits will be the huge time saved for cars travelling through the traffic network. The success of using SensorGrid with Q-learning algorithm on this traffic problem allows more sophisticated applications to be developed in the future.