Observability analysis for improved space object characterization

An information-driven estimator is developed to allow for more accurate estimation of the orbit and appropriate parameters, such as area-to-mass ratio and attitude, of a space object. From the information dilution theorem, when additional states are added to a filter with the number of observations held constant, the uncertainty in each state will increase. Likewise, failure to compensate for uncertainty in system states and/or parameters requires process noise compensation, which increases the uncertainty of the estimate. The system observability can be used to determine when enough information exists to estimate additional states and which states can be appropriately estimated at that time. Tracking examplesofspace objectsinmultiple orbit regimes from angles-only data are presented, examining the estimation accuracy and uncertainty of the orbital states, both with and without estimation of area-to-mass ratio. Light-curve measurements of a faceted vehicle are added, and estimation of orbit and attitude is considered. It is shown that the proposed adaptive approach based on the observability and assessment of measurement information content performs better over the entire time period than a single filter with a constant number of states.