A TWO-LEVEL OPTIMAL ESTIMATOR FOR ATTITUDE DETERMINATION USING GPS MEASUREMENTS

摘要

This paper focuses on applying the two-level optimal estimator to the problem of attitude determination using GPS. The derivations are based on the two kinds of criteria: the traditional one using a set of positive scalars as the weighted factors, and another one employing the inverse of covariance of measurements as the weighted matrix. It is shown that the configuration to formulate Wahba's problem as the criterion of the second level is numerically efficient. The solution of the two-level optimal estimator turns out to be optimal under the so-called balanced condition. The experiment results show the efficacy of the two-level optimal estimation method.