Objectives: To develop methods and techniques to reliably collect sensor data from diverse sources and to capture the context of a situation, to develop methods to analyze context and assess a situation, to develop and assess programming abstractions and a framework for correct design of distributed processing algorithms.
This work package develops techniques that detect events as well as abnormal occurrences and recognize out of the order activities in public spaces. It investigates new programming abstractions such as rule-based programming techniques to program a scalable network with hundreds of sensors e.g., as found in LANE type networks.
For CLOUD networks this WP investigates data fusion by ensembles of smart sensor phones. An ensemble is an opportunistic cluster of smart sensor phones, with a collective set of sensors large enough to detect context and situations. Phones in an ensemble may have different kinds of sensor, but by sharing and fusion of sensor data, phones complement each other. Ensembles are highly dynamic and will search for and include phones until the set of collective sensors is sufficient to assess a situation. Mobile phone ensembles (CLOUD) are augmented by sensors in the infrastructure around, e.g. street lighting (LANE).
WP4 uses the results from WP1 and 3 and its results are input for WP1 and 5. Task T4.1 Distributed data fusion-based event detection in ensembles of smart sensor phones (UT/PS) Task T4.2 Distributed signal processing and control for situation assessment (TU/ES) Task T4.3 Programming framework (TU/SAN)