Center for Embedded Network Sensing

Monitoring of environmental phenomena with embedded networked sensing confronts the challenges of both unpredictable variability in the spatial distribution of phenomena, coupled with demands for a high spatial sampling rate. For example, high fidelity sensing of light intensity under a canopy of tree may require spatial sampling rate of 10 samples/m2 over a transect exceeding 1000 m2. Clearly, adequate sampling coverage of such a transect requires an impractically large number of sensing nodes. We propose here a multi-scale scheme of sampling such environments, using Networked Infomechanical System (NIMS). In this multi-scale paradigm, we use imager for high resolution and low fidelity sensing that provides us with high-level, global view of the environment. This high resolution information is then used for actuated guided sampling using mobile sensors attached with NIMS that provide low resolution, high fidelity sensing information for regions of interest.