Ground tests of magnetometer-based autonomous navigation (MAGNAV) for low-earth-orbiting spacecraft

Abstract

Actual spacecraft data have been used to test two filters that rely solely on magnetometer measurements to determine spacecraft ephemeris. This work proves the potential of a novel autonomous orbit determination scheme. One of the filters is an extended Kalman-type filter. It estimates a position and velocity, and it uses equations of motion to propagate these estimates. The other filter is a batch least-squares filter. It estimates static Keplerian parameters along with magnetometer biases. Tests of these filters involve filtering of magnetometer data followed by comparison of estimated position with position derived from ground-based tracking. Both filters perform well on the MAGSAT spacecraft, achieving accuracies from 4 to 8 km. Tests of the batch filter on the DE-2 spacecraft yield somewhat larger position uncertainties, 17-18 km. Batch filter tests on the LACE spacecraft yield much larger position errors, 120 km. This poor LACE performance can be attributed partly to solar-array-g enerated low-frequency noise.