Nighttime Composite Moonlight Fields Enable Orientation Determination in GNSS-Denial

In global navigation satellite system (GNSS) denied and low-light nighttime environments, it is challenging to achieve accurate and robust attitude determination of uncrewed systems. Inspired by nocturnal insects, this article presents an autonomous navigation approach that relies on the composite light field (CLF), comprising the lunar-reflected radiance field (LRF) and the polarized light field (PLF). Subsequently, a CLF attitude and heading reference system (CLF-AHRS) is developed. The system enables the synchronous acquisition of both LRF and PLF, while a two-level robust fusion scheme is employed. By incorporating horizontal attitude information from the inclinometer, the method ensures reliable nighttime navigation, especially in cases where the LRF is subject to occlusion. Both simulations and outdoor experiments demonstrate that the proposed approach effectively addresses the limitations of LRF when affected by occlusion interference. Compared to the PLF-based method, it achieves a significant improvement in accuracy under both physical occlusion and cloud interference. Furthermore, the system inherently eliminates volumetric redundancy and systemic errors that are typically induced by a distributed multioptical sensor configuration. These developments substantially enhance the robustness and adaptability of navigation for uncrewed systems in complex and constrained nighttime environments.