Advanced Monocular Passive Ranging (A-MPR)

"The proposal addresses an advanced method of Monocular Passive Ranging (MPR). MPR is a differential absorption technique to determine the range to an emissive target without the use of active illumination (RADAR or LIDAR). Classical MPR uses sensormeasurements in two discrete bands (an atmospheric window and an absorption band), assumes the target source radiance ratio and atmospheric conditions, and then solves directly for range. This method suffers from the fact that incorrect parameterassumptions produce poor range retrievals.We introduce a novel computational paradigm that can take advantage of the target spectral signatures measured in multiple bands, and in which physics based model (MODTRAN in this case) predictions are combined with sensor measurements to retrieve theBayesian best estimate of the parameters. The algorithm, Simultaneous Constraint Optimization for Parameter Extraction (SCOPE), is based on a constraint optimization of an objective function that expresses the best estimates weighted by the covariance ofthe uncertainties in the initial guess of the parameters, uncertainties in the sensor measurements, and model uncertainties. The primary constraint is an analytical expression which is the Taylor expansion of the complete MODTRAN model. The Jacobianmatrix for the expansion is obtained from the Automatic Differentiation augmented MODTRAN code. The proposed computational paradigm, SCOPE, lends itself to compu