Sensor networks for real-time vehicle tracking on road networks

Due to the continuously increasing mobility demand the efficiency of the existing road infrastructure should be improved. Advanced control and management techniques offer promising, cost effective and environmentally friendly solutions to the problem. As traffic and vehicle control/management applications become more advanced they require more and more detailed and accurate real-time observations about the surrounding relevant world. The paper describes a road infrastructure based sensing and data interpretation system, which is scalable, robust and delivers the motion state estimation of the individual vehicles in real-time. The system enables a wide spectrum of intelligent transport system functionalities ranging from traffic management to fully autonomous driving. Distinguishing features of the solution that it does not require instrumentation on the vehicles to become detected and it is insensitive to ambient environmental conditions (weather, light, etc.). The paper describes the event driven object tracking algorithm, and its "mapping" to distributed computing platform resulting in an inherently robust implementation.