Adaptive Estimation of a Powered Spacecraft

An adaptive estimator is presented and performance is demonstrated which successfully tracks a powered spacecraft and estimates the spacecraft state. The system dynamics are inherently nonlinear due to the presence of gravity. The measurement model, with measurements of line-of-sight range, azimuth, and elevation, is also nonlinear. The estimator is derived from an ARMAX formulation which assumes a linearized, time-varying dynamical system with a deterministic acceleration. In this paper, two algorithms are derived. The first algorithm is an adaptive batch estimator which uses the innovations to estimate a piecewise constant acceleration approximation to the real system dynamics. The second algorithm is an Iterated Least Squares Polynomial time estimator which models the third time derivative (jerk) as a constant. Finally the algorithms are used in a tracking scenario to provide a performance comparison.