GNSS/INS Navigation System Based on Lie Group Right Invariant Kalman Filter

Invariant Kalman filter is a novel Lie group-based method to solve the state estimation challenges of unmanned systems. It has garnered widespread attention because its error dynamics are not dependent on the current state, enhancing its filter consistency. In this paper, the GNSS/INS integrated navigation system is built on the right invariant Kalman filter, and its right invariant error propagation model and observation equation are derived. We further used the real-flight data of the fixed-wing UAV for offline validation. Compared with the nonlinear complementary filter algorithm based on manifold (NCF) and the standard error state EKF algorithm(ES-EKF), our method ensures a notably swifter pose convergence and superior accuracy.