Batch Sequential Estimation with Non-Uniform Measurements and Non-Stationary Noise

Sequential estimation using the traditional discrete Kalman filter typically
assumes the measurement time and state update time are coincident. This is often
a poor assumption in realistic measurement scenarios where the data can be
received from multiple sources at differing times. This paper develops the
necessary algorithm adjustments needed for the Kalman filter to readily
process measurement data that arrive at varying times and with non-stationary
noise. The algorithm is applied to a relevant problem of orbit determination using
one-way uplink radiometric tracking of a spacecraft (in the present case a Mars
orbiter).