High-dimensional system error model and real-time estimation scheme for GNSS/INS integrated system

Abstract This paper establishes an advanced high-dimensional system error model with 42 Kalman filter states for a GNSS/INS integrated system, which sufficiently considers the influence of multiple INS crucial error factors according to the remote sensing survey requirement. Moreover, to overcome the heavy computational burden imposed by high-dimensional matrix operation, an optimal state estimation scheme with high-efficiency Kalman filter is proposed by comprehensively making use of matrix sparsity and symmetry, which retains the estimating optimality with low computation complexity. It is demonstrated through flight experiment that the proposed 42-dimensional error model offers better estimation and navigation precision than the other models, and the proposed real-time estimation scheme can effectively reduce the computation and memory usage to improve the overall performance of GNSS/INS.