Enhancing Attitude Availability in Star-Depleted Cases: An Inertial/Star Sensor Fusion Method

This article presents a robust fusion method to address stellar observation scarcity in satellite attitude determination. Conventional systems fail with fewer than three navigation stars, jeopardizing missions requiring precise pointing. The proposed tightly coupled architecture embeds a hemispherical resonator gyroscope-based inertial reference within the star sensor's processing chain, which enables predictive star tracking during scarcity events. By reformulating an observation model and implementing an inertial-aided matching algorithm, the tightly coupled system maintains resilient performance across varying stellar visibility. A tailor-made Kalman filter explicitly formulated around the novel observation model is designed to reliably achieve the process of state prediction, update, and implementation. Experimental results demonstrate that the proposed method not only achieves accuracy better than one arc-minute under optimal conditions but sustains reliable solutions with just one or two stars, which eliminates the traditional three-star threshold. This capability significantly enhances mission reliability in star-depleted regions.