Magnetometer-Based Attitude Determination Using Common-Frame Filtering with Unobservable Parameter Consideration

In this work, a common-frame state estimation routine is adapted to accommodate unobservable parametric quantities. Using the traditional approach taken in consider or Schmidt Kalman Filtering, pseudo-states are augmented in filter design to accommodate quantities that are undesirable to be updated via measurements. The filter is derived to provide frame consistency for both traditional and pseudo-states. When applied to magnetometer attitude estimation of spacecraft, simulation results using the consider common-frame filter yield more consistent state estimates with the added benefit of faster state convergence as compared to the standard consider filter. The resulting consider common-frame filter has wide utility to nonlinear systems with lightly observable or unobservable parameters while maintaining a common-frame formulation.