Calibration of Space-Based Sensors and Unknown Target Signature Separation

Applied Optimization, Inc. proposes a three-step approach for identifying the unknown physical characteristics of a target satellite based on its multi-spectral point image. In the first step, a method is proposed for calibrating space-based passivesensors using data from tracking a Navstar satellite. The specific data to be used was obtained by an experiment aboard the MSX satellite. There are two advantages to this approach, namely; the Navstar position is known precisely, and a constellation oftwenty-four satellites is always available for tracking. The second step considers that Navstar characteristics were unknown and represents them as a Boolean of geometry primitives and a set of surface properties. Its point image is expressed as a sum ofseveral contributions that arise as the spacecraft swivels to manage the pointing of its antenna cluster and solar panels. Then the visual and infrared spectrum data is fused to solve for unknown characteristics. The third stage considers that the targetcould be represented as a Boolean of unknown geometry primitives, which can be crossed-over and mutated to match the multi-spectral point image. A Genetic Algorithm is proposed to create the different possibilities and decode the components in the unknowntarget signature.DoD Benefits: [1] A process for calibration of passive sensors using mono tracking data. [2] Reduction of time between detection and decoding of an unknown target. Applications: [DoD] Separation of unknown signature for ballistic missiletargets. [Commercial] Analysis of Measurements for Highway Traffic Signature Separation.