Research On The Target Tracking Algorithm Of Combined BDS/SINS Navigation System Based On Cubature Kalman Filter

Nowadays,satellite navigation systems are used in various fields and their position is crucial,so each country wants to develop a satellite navigation system with fully independent intellectual property rights.In recent years,China’s Beidou Navigation System has developed rapidly and has become a mature satellite navigation system in the world.Since the performance and stability of a single navigation system can no longer meet the demand for navigation in various industries,some scholars have proposed the concept of combined navigation.The main idea is to use the respective advantages of different navigation systems to both complement each other’s functions and improve the accuracy and stability of navigation.In this thesis,we focus on the improved data fusion algorithm for the combined navigation system and the study is mainly conducted in the context of a tightly combined navigation system of BDS and Strap-down Inertial Navigation System.In this thesis,the traditional data fusion methods,Kalman Filter,Extended Kalman Filtering,Unsecented Kalman Filtering and Cubature Kalman Filter,are investigated for the strong nonlinear characteristics of the SINS.The results of the simulations are compared and analyzed.The experimental results show that the CKF algorithm has the best overall filtering effect in the BDS/SINS tight combination navigation system,so it is chosen as the base algorithm in this paper.For the CKF algorithm in the filtering process,it needs to know the a priori information of the system noise and the statistical characteristics of the measurement noise,otherwise it will lead to the degradation of the filtering accuracy or even the filtering scattering,but in the actual environment these information may be unknown.Based on this problem,this thesis proposes an adaptive filtering algorithm for noise estimation through the a priori information of the unknown noise,by introducing a constant noise valuator and time-varying noise valuator to predict the statistical properties of the unknown noise in real time.Simulation experiments are conducted for these two noise valuator algorithms respectively,and the experimental results show that the noise valuator has a good effect on the real-time tracking of noise,and the ACKF algorithm with noise valuator has an adaptive capability in the face of unknown noise a priori information,and it is possible to do without the known noise a priori information before filtering.For the BDS receiver in accepting the return signal from the Bei Dou satellite,it is likely to receive anomalous values by interference,which will deteriorate the filtering accuracy and stability.In the filtering process,due to the increase of the number of filtering iterations,the error covariance matrix in the noise statistical characteristics may appear non-positive definite,resulting in the inability to perform the Corsi decomposition filtering interruption.Therefore,it is proposed to first preprocess the initial observations with mean filtering so that the influence of outliers on the filtering is reduced;then use Singular Value Decomposition instead of the Corsi decomposition to avoid the non-positive-definite filtering interruption of the error covariance matrix,because the SVD decomposition can decompose the non-positive-definite matrix;finally construct the anti-difference factor using the improved IGGIII weight function to reduce the influence of anomalous observations on the filtering again.The above algorithm is proposed on the premise that it is implemented in a BDS/SINS tight combination navigation system.In order to verify the effectiveness of the proposed algorithm,the framework structure of the tight combined navigation system is designed,the simulation environment required for the experiment is built,and the above proposed improved algorithm is used in the experimental simulation platform.Compared with the traditional CKF filtering algorithm,the experimental results show that the SVD-based anti-difference adaptive CKF algorithm has better adaptivity to the unknown noise a priori information and better anti-difference capability to the abruptly changed observations in the observed information.The proposed algorithm is proved to be a feasible and effective filtering algorithm in the tight combined navigation system.