NESL Technical Report #: 2006-7-1
Abstract: This report was a precursor to Chapter 4 in Laura Balzano's masters thesis.
The purpose of this project is to explore the method of the Ensemble Kalman Filter as it could be used in sensor networks. In particular, we examine how the Ensemble Kalman methods work under faulty data that is commonly found it sensor networks.
Sensor networks are envisioned to have large numbers of very small, very inexpensive devices. We hope to more than offset any detrimental affect from noisy, faulty data with the high temporal and spatial resolution brought on my increasingly many devices. In light of these characteristics, we will need statistical tools to help us leverage the vast amounts of sensor data without getting tripped up by the low quality of any individual sensor’s measurements. The Ensemble Kalman Filter is a point-mass filter which tracks the pdf of the state of a dynamical system by using a Monte Carlo technique. In this report, we discuss the effectiveness of the Ensemble Kalman Smoother when estimating these parameters on an autoregressive model of decaying soil moisture.
NESL Document?: Yes
Document category: Report