Dynamical Model Compensation for Near-Earth Satellite Orbit Determination
KENNETH A. MYERS, Byron D. Tapley · AIAA Journal · 1975
Operational requirements in modern space applications often demand orbit determination accuracies which are limited by fundamental mathematical and computational restrictions. It is shown in realistic computer simulation studies how these difficulties can be alleviated for typical near-Earth satellites by employing dynamical model compensation (DMC) and accurate observations in the extended Kalman filter. Unmodeled and unknown accelerations affecting the motion of the satellite are effectively compensated by treating them as a firstorder, Gauss-Markov stochastic process. Although conventional state noise compensation (SNC) can provide satisfactory results for many applications, the DMC method offers a significant increase in estimation accuracy. Numerical behavior of the DMC structure is summarized as a function of a priori statistical parameters to aid in filter design analyses for operational applications.