Accession Number : AD1001955


Title :   On-line Flagging of Anomalies and Adaptive Sequential Hypothesis Testing for Fine-feature Characterization of Geosynchronous Satellites


Descriptive Note : Conference Paper


Corporate Author : Air Force Research Laboratory Wright Patterson AFB United States


Personal Author(s) : Chaudhary,Anil ; Payne,Tamara ; Kinateder,Kimberly ; Dao,Phan ; Beecher,Elizabeth ; Boone,Derek ; Elliott,Brittany ; Billing,Hailey


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/1001955.pdf


Report Date : 18 Oct 2015


Pagination or Media Count : 67


Abstract : The objective of on-line flagging in this paper is to perform an interactive assessment of geosynchronous satellites anomalies such as cross-tagging of satellites in a cluster, solar panel offset change, etc. This assessment will utilize a Bayesian belief propagation procedure and will include an automated update of the baseline signature data for the satellite, while accounting for the seasonal changes. Its purpose is to enable an ongoing, automated assessment of satellite behavior through its life cycle using the photometry data collected during the synoptic search performed by a ground or space-based sensor as a part of its metrics mission. The change in the satellite features will be reported along with the probabilities of type I and type II errors. The objective of adaptive sequential hypothesis testing in this paper is to define future sensor tasking for the purpose of characterization of fine features of the satellite. The tasking will be designed in order to maximize new information with the least number of photometry data points to be collected during the synoptic search by a ground or space-based sensor. Its calculation is based on the utilization of information entropy techniques. The tasking is defined by considering a sequence of hypotheses in regard to the fine features of the satellite. The optimal observation conditions are then ordered in order to maximize new information about a chosen fine feature. The combined objective of on-line flagging and adaptive sequential hypothesis testing is to progressively discover new information about the features of geosynchronous satellites by leveraging the regular but sparse cadence of data collection during the synoptic search performed by a ground or space-based sensor.


Descriptors :   geosynchronous satellites , brightness , clustering , brownian motion , sequential analysis , test methods


Distribution Statement : APPROVED FOR PUBLIC RELEASE