Prof. Paul Racette
NASA Goddard Space Flight Center
December 15, 3:30 PM
Thirkield Hall (Physics), room 103
Ensemble Detection and Analysis
This presentation will describe Ensemble Detection and Analysis as a novel approach to analyzing and modeling non stationary processes using methods derived from the study of the performance of radiometer system calibration architectures. Everything changes across some temporal or spatial scale. The lack of well-developed techniques for modeling changing statistical moments in our observations has stymied the application of stochastic process theory for many scientific and engineering applications. Nonlinear effects of the observation methodology, i.e. the role of the observer, is one of the most perplexing aspects to modeling non stationary processes. For example, such nonlinear effects are problematic when averaging high resolution radar data to match courser resolution radiometer data in combined retrieval algorithms. These limitations were encountered when modeling temporal effects of calibration frequency on the performance of a radiometer with non stationary receiver fluctuations. A microwave radiometer is frequently calibrated to correct for fluctuations in the receiver. A radiometer typically samples a set of stable calibration noise references from which the receiver response is estimated. Algorithms are usually applied to suppress receiver fluctuations from the estimates of the measurand. Analysis has shown that algorithms designed to accentuate temporal effects in the receiver response yield information about the non-stationary properties of the receiver fluctuations. Ensemble Detection and Analysis extends this concept to the study of non stationary signals as a form of noise assisted data analysis. The presentation will conclude with musings on the ontology of a new Observation Theory.
Refreshments will be served at 3:15 pm