Sensor data collection in a transport monitoring environment
- Speaker: Dr Niki Trigoni (Birkbeck, University of London)
- Host: Norman Paton
- 13th December 2006 at 14:15 in 1.5
Technological advances in sensor technology offer unprecedented opportunities for fine-grained and continuous monitoring of road traffic. The availability of traffic information is paramount to preventing or handling traffic jams, detecting accidents and automating road maintenance. In this talk, I shall first overview the main research challenges investigated in the context of the TIME-EACM project (Transport Information Monitoring Environment: Event Architecture and Context Management), ranging from low-level sensing and data collection, to middleware and access control services for querying traffic data. I shall then focus on the problem of collecting traffic data from the sensor nodes and disseminating it to the middleware layer. By exploiting the nature of spatio-temporal correlations in an urban traffic scenario, it is possible to compress data within the sensor network and incur significant communication savings, whilst adhering to user-defined accuracy requirements. Furthermore, spatio-temporal correlations are key to determining how to place sensor nodes in an urban setting and to minimizing the initial cost of sensor deployment. Preliminary results on the performance of data collection and sensor placement strategies will be presented, using real SCOOT traffic data generated in the city of Cambridge.