High-Accuracy Double Grid Relative Navigation Framework for UAV Swarm in GNSS Challenging Environment

Relative navigation technology is crucial for cooperative operations of unmanned aerial vehicle (UAV) swarm in global navigation satellite system (GNSS) challenging environment. There are two critical functionalities of relative navigation technology: accurate relative navigation estimation in the scenarios of complete GNSS denial and absolute navigation information diffusion when the GNSS become available for partial UAVs. Existing relative navigation algorithms cannot achieve good performance in terms of the two functionalities at acceptable communication and computational costs. This paper proposes a high-accuracy double grid relative navigation framework. The more accurate grid frame drift formulas and double grid error propagation models are constructed, which improves the ability of accurate relative navigation estimation. The grid offset measurement model is derived to exploit the relationship between the state estimates in double grid frames, which enhances the performance of absolute navigation information diffusion. A general source selection strategy for arbitrary measurement models is proposed to suppress the estimation overconfidence, and its functionalities are analyzed and discussed. The proposed theoretical framework is implemented in the application of datalink-based UAV swarm relative navigation. Simulation and experimental tests have been carried out, which validate the effectiveness and superiority of the proposed framework.