Non-Linear Optimization Applied to Angle-of-Arrival Satellite-Based Geolocation with Correlated Measurements

A common remote sensing application is producing geolocation estimates for an object of interest from multiple sensor platforms. Geolocation estimates are desired to help improve situational awareness when dealing with space objects that do not actively broadcast their location. A depiction of the error parameters are calculated in conjunction with the positional estimates. Problems occur when multiple measurements from a single sensor are used to estimate a location due to correlations in sensor error. A non-linear optimization approach is presented for determining geolocation estimates and their associated error parameters. The error parameters directly reflect the error present on the individual measurements used to produce the position estimates. Correlations in errors are dealt with by augmenting the non-linear optimization with a covariance intersection algorithm. Finally, the ability to account for correlated errors within the optimization algorithm is analyzed using Monte-Carlo simulations. The ability to describe an objects location with a given confidence helps aid in the analysis of the system at large