A Technical Seminar presented by the Washington/Northern Virginia Chapter

of the IEEE Geoscience and Remote Sensing Society

 

Novel Self-Adpative Application Strategies using Smart Satellites

 

Dr.  Richard B. Gomez

School of Computational Sciences

George Mason University

Fairfax, VA

rgomez@gmu.edu

 

 

Abstract:  This talk presents self-adaptive application strategies to operationally process and exploit spectral information on-board smart remote sensing satellites in real time.  The goal is to greatly increase onboard processing capabilities and reduce data volumes for hyperspectral systems by incorporating this intelligent-based approach.  The approach focuses on the future development of hardware and software methods using an architecture based on Reconfigurable Computers that work more like the human brain and body.  These smart satellites will be self-sensing, self-repairing, and self-adaptive.  If something is not exactly right, the smart satellite will not need a signal from ground control.  The spacecraft itself decides what's best and then take the necessary action to address the problem.  It’s a revolutionary means to process and exploit on board, huge amounts of hyperspectral data without the need to have an expert spectral analyst in the loop.  Special attention will be paid to spectral libraries and novel usage of exemplar spectra.  This involves techniques that are beyond the forefront of current database research.  NASA has recognized the value of using a hyperspectral sensor flying with or slightly behind of the imaging sensor platform to identify features of interest to be processed.  The hyperspectral sensor is used to identify the presence of a feature and verify its composition.  The employment of a near real time feature recognition system on the hyperspectral sensing satellite is invaluable.  We envision that a reconfigurable architecture incorporating evolutionary algorithms will be developed knowing that self-adaptive satellites may be the key to success.