ERROR ANALYSIS FOR THE STANFORD RELATIVITY GYROSCOPE EXPERIMENT

This thesis discusses an error analysis performed for the Stanford Relativity Gyroscope Experiment. The experiment is designed to measure any relativistic precession of a gyroscope carried in an earth-orbiting satellite to an accuracy of (+OR-)1 milliarcsec/yr. Einstein's General Theory of Relativity predicts two precessions: A "geodetic" precession due to the orbital motion of the gyroscope about the earth (6.9 arcsec/yr) and a "motional" precession due to the interaction of the earth's rotation with the spinning gyro (0.044 arcsec/yr).; There are four superconducting electrically suspended spherical gyroscopes on the satellite. The direction of their spin axes is measured by means of SQUID magnetometers and compared with an inertial reference. The plan is to use the star Rigel as the "inertial" reference. Its direction is measured using a telescope on the satellite.; The error in determining the relativistic precessions was calculated using a Kalman filter convariance analysis with a realistic error model. An averaging technique was used to reduce the amount of computer time needed to perform the convariance analysis by a factor of (TURN) 1000.; Studies show that a slightly off-polar orbit is better than a polar orbit for determining the "motional" precession, i.e., depending on the a priori uncertainty in Rigel's proper motion either a 70 deg orbit or 86.25 deg orbit is best for determining the motional precession. SQUID magnetometer noise turns out to be the dominant measurement error source. Depending on its magnitude, the a priori proper motion uncertainty may be the dominant error source in the experiment. If Rigel's proper motion is known exactly, the motional precession can be measured to an accuracy of 0.52 milliarcsec/yr in an 86.25 deg orbit.