Bias-Aware Tracking Window Optimization for Autonomous Navigation of Cislunar Satellite Formations

Abstract

As interest in lunar exploration grows, efficient and cost-effective navigation methods are becoming increasingly important. Traditionally, ground-based radiometric tracking has been the primary approach for spacecraft navigation, particularly for missions beyond Earth. In contrast, Satellite-to-Satellite Tracking (SST) enables onboard Orbit Determination (OD) without continuous ground support. This study focuses on optimizing SST periods for a cislunar satellite formation to improve autonomous navigation performance. The timing of tracking windows is critical, as selecting the most informative observation periods can significantly enhance OD accuracy and reduce station-keeping costs. Systematic measurement biases present additional challenges, making careful scheduling essential to mitigate their effects. To address this, we employ Particle Swarm Optimization (PSO) to determine effective tracking schedules while accounting for bias constraints. The results show that this approach yields near-optimal schedules that outperform uniform, evenly spaced observation strategies with respect to the objective function, defined as a weighted combination of OD error and station-keeping costs, even under non-optimal parameter settings. The optimized schedules are irregularly distributed over the mission timeline, reflecting adaptation to time-varying geometries and uncertainties. This approach improves autonomous navigation accuracy, reduces station-keeping costs, and supports the cost-effective design of cislunar satellite formations, thereby improving the feasibility of satellite operations beyond Earth.