Moving Sensors

This collaboration project with database lab will address the following issues:

  • Multilayer framework for processing data streams produced by moving sensors

This multilayer framework consumes streams of data and construct their multi level, spatio-temporal representation for further manipulation and processing.

This proposed framework provides the following key benefits:

  1. It provides physical data independency meaning one may change the physical characteristic of the sensors without re-writing the software that implements the spatio-temporal layers

  2. This framework separates data from process which provides both modularity and extensibility

  • Describing the techniques to acquire data from moving sensors and proposing a novel and challenging way to deal with noise in the data stream

  • Investigating the techniques that describe the essence of an activity and constructing the Envelop of Limits(EoL) based on repetition of a task and some domain specific knowledge


Research Agenda

  The main focus of our research has been on extending our "Multilayer Framework" for Gesture Recognition. Since the gesture recognition techniques often suffer from being highly device-dependent and hard to extend, we have introduced a "Multilayer" approach that can address these two aforementioned issues.

The immediate research agenda of this proposal has two aspects:

  1. Completing the design of the multi-layer framework by completing its spatio-temporal component
  2. Evaluating  the extensibility of the framework






Sponsors :

IDM Program