Annual Report 2008.pdf - SAMSI

The main goal of a data collection protocol for sensor networks is to keep the network’s database
updated while saving the nodes’ energy as much as possible. To achieve this goal without
continuous reporting, the data suppression is a key strategy. The basic idea behind of data
suppression schemes is to send data to the base station only when the nodes’ readings are
different from what both nodes and base station expect. Data suppression schemes can be
sensitive to aberrant readings, since these outlying observations mean a change in the expected
behavior for the readings sequence. In this paper, we present a temporal suppression scheme that
is robust to aberrant readings. We propose to use a technique to detect outliers from a time series
as the basis of a temporal suppression scheme named TS-SOUND (Temporal Suppression by
Statistical OUtlier Notice and Detection). TS-SOUND detects outliers in the sequence of sensor
readings and sends data to the base station. Outliers can suggest a distribution change-point or be
an aberrant reading. Then, we have adapted TS-SOUND to avoid detecting aberrant readings
and, even this filter fails, TS-SOUND does not send the aberrant reading to the base station.
Experiments with real and simulated data have shown the TS-SOUND scheme has suppression
rates comparable and even greater than the rates of temporal suppression schemes proposed in
the literature. Furthermore, it keeps the prediction errors at acceptable levels, if we consider the
complexity of the data collection using a sensor network.
Harsh Singhal
University of Michigan
singhal@umich.edu
“Optimal Experiment Design for State Space Models with Application to Sampled Network
Data”
We introduce a linear state space model for the estimation of network traffic flow volumes from
sampled data. Further, we formulate the problem of obtaining optimal sampling rates under
router resource constraints as an experiment design problem. Theoretically it corresponds to the
problem of optimal design for estimation of conditional means for state space models. We
present the associated convex programs for a simple approach to it and propose several exciting
problems in this area. The usefulness of the approach in the context of network monitoring is
illustrated through an extensive numerical study.
Zhengyuan Zhu
University of North Carolina -Chapel Hill
zhuz@email.unc.edu
“Optimal Network Design for Detecting Regional Trend in PM”
G. Beate Zimmer
Texas A&M University-Corpus Chris
beate.zimmer@tamucc.edu
“Texas Coastal Ocean Observation Network: Some Applications of the Wealth of Data”