Application of Kalman filter in GPS-based attitude determination

In this paper, several issues in applying nonlinear Kalman filter to GPS-based attitude determination are addressed. The governing equation that characterizes the attitude information and GPS measurements is developed. A nonlinear Kalman filter is then formulated to estimate attitude, integer ambiguity, and other relevant parameters. To account for the integer constraints and over-parametrization on the attitude, several schemes are proposed to augment standard Kalman filter iterations. An experiment using a rotating mechanism to emulate a spin-stabilized satellite and two carrier phase GPS receivers to provide measurement data is conducted in this work to assess the feasibility of the proposed algorithm.