Robust GNSS/INS Integrated Navigation Method Based on Variational Bayesian and Adaptive Maximum Correntropy
Tharindu Gamage, Minbo Yang, Hao Wu, Chenchen Jiang, Xiaoran Cao · 2025
A novel robust GNSS/INS integrated navigation algorithm is proposed that combines variational Bayesian (VB) adaptive filtering with adaptive maximum correntropy-based robust estimation (VBAMCC) to address two challenges in GNSS/INS systems such that adaptation to time-varying noise and heavy-tailed noise rejection. The VBAMCC approach introduces the variational Bayesian estimation to adapt noise covariances in real-time and uses a mixed kernel correntropy criterion to down-weight outlier effects. The simulation results show that it effectively handles both measurement outliers and changing noise conditions, outperforming the traditional algorithms.