Improved innovation-based adaptive estimation for measurement noise uncertainty in SINS/GNSS integration system

The Kalman filter (KF) is the most common method for the estimation problems of the integrated SINS/GNSS system, but its performance depends on the correct a priori knowledge of model dynamics and noise statistics. The GNSS measurement noise uncertainties will degrade the performance of the KF for the fixed measurement noise covariance matrix. To fulfill the accuracy requirements of the dynamic system, an improved innovation-based adaptive estimation (IAE) algorithm is proposed. Based on the IAE principle, a regulatory factor is introduced into the calculation of the gain matrix to solve the singular value problem during the matrix inverse operation, and cut down the estimation errors caused by measurement noise uncertainties. The performance of the proposed algorithm is evaluated by the Monte-Carlo simulations in the SINS/GNSS integration system and significant improvements on the filter performance have been achieved.