Sensor Calibration as an Application of Optimal Sequential Estimation toward Maintaining the Space Object Catalog

As space-object catalogs move from being maintained by general perturbations to special perturbations, and as the number of detectable objects in space increase, sequential (recurrent) estimators have some significant advantages over batch processing that should not be overlooked. A primary advantage of a sequential estimator is that the time evolution of the satellite state error covariance can be more accurately determined. The availability of accurate uncertainty estimates makes other types of operational analyses easier, including sensor calibration. A sequential estimator can also provide estimates in near real-time, which can be very important for sensor calibration. In this paper, an optimal sequential estimator is applied, via simulation, to the problem of detecting and estimating large changes in sensor performance closer to real time. By using statistical significance tests on the estimated biases, a possibility exists to detect substantial changes in sensor performance more quickly and accurately.