A Risk-Aware Robust Navigation Framework for UAVs in GNSS-Degraded Low-Altitude Environments

To address the critical issues of geometric ill-conditioning and non-line-of-sight (NLoS) interference faced by broadcast radio positioning systems in long-distance transmission (≥200 km) and low-altitude flight scenarios (1000 m~3000 m), this paper proposes a Differential and Robust Positioning method for Airborne Platforms (DPAP). By integrating radio differential positioning technology, this method improves the single-point positioning algorithm through grid search and iteratively reweighted least squares to mitigate geometric ill-conditioning and numerical instability in low-altitude environments. Furthermore, a passive differential positioning approach is introduced to eliminate common errors using neighboring reference stations. Finally, a scenario-aware safe fusion strategy ensures that the fused solution is never inferior to the optimal sub-solution under any circumstances. Simulation results demonstrate that, under conditions involving six ground stations, user-to-station distances of no less than 200 km, and 15% of links experiencing NLoS propagation, the differential and robust positioning method achieves a positioning accuracy of 0.588 m RMS. This represents an improvement of approximately one order of magnitude compared to RSPP (12.304 m), and outperforms traditional Huber M-estimation (0.678 m) and elevation-weighted least squares methods (1.462 m).