Unresolved Resident Space Object Characterization Using Time-Frequency Analysis

Applied Optimization, Inc. [AO] proposes to demonstrate full feature characterization of unresolved RSOs by performing spectral unmixing and time-frequency analysis of its time and spectrally resolved signature in the visible or infrared spectrum. The characterization will be performed step-by-step by following a taxonomy of feature extraction. Materials identification will be performed via spectral unmixing. RSO typing will be performed on the basis of materials present in the signature. Time frequency analysis of temporal behavior of material fractional abundances will be utilized to characterize the RSO mechanics. The algorithms will be implemented as reusable software components in a tool kit format. A Bayesian belief propagation technique will be used to navigate the RSO characterization taxonomy and to assign a measure of goodness to the results. This will be accompanied by a tool for visualization of the RSO feature extraction sequence. Quality historical data will be utilized to generate a RSO time-frequency behavior catalog which can be used for correlating or identifying unknown objects. The effectiveness of the algorithms and the RSO data catalog will be verified by newly collecting and analyzing time and spectrally resolved RSO signature data in the visible and infrared spectra.