Research On Integrated Navigation Filtering Algorithm Based On CKF

The integrated navigation system has become a hot spot in the field of navigation since its birth,and the integrated navigation system of GNSS and INS is the hotspot in the field of integrated navigation system.The integrated navigation system has the following characteristics:(1)For GNSS receivers,the receiver tracking and capture abilities can be enhanced by inertial navigation assistance.(2)For inertial navigation,it not only can have the short-term high precision of inertial navigation system and can work with itself of no external interference,but also can suppress inertial navigation over time accumulation error with GNSS,then enhance the use of the navigation system alone,the reliability and accuracy.In this paper,with the continuous development of the missile navigation system,how to reduce its cost while improving the missile precision guidance is the hotpoint and difficulty of the research.Therefore,the current missile navigation system mainly studies the integrated navigation system based on low precision MEMS INS and GPS,and it is necessary to improve the combination of MEMS INS and GPS The need for a integration of more intensive ultra-tightly coupled integrated navigation mode.At present at home and abroad on the ultra-tightly coupled integrated navigation system has done a lot of research and achieved some results,but the ultra-tightly coupled integrated navigation systems still have the following questions:(1)System complexity,the system model is especially complicated nonlinear system.The traditional Kalman filter can not deal with such a complex nonlinearity,especially the traditional I/Q signal of the receiver is used as the centralized ultra-tightly coupled integrated navigation system of measurement information.(2)The existing GNSS/INS ultra-tightly coupled integrated navigation system model is relatively simple,and its statistical characteristics of the noise is determined,when the system is used in complex environments,such as weak signal strong interference environment,the use of this single model can not accurately describe the system,and will cause navigation information a larger error.Therefore,this paper based on Missile GPS/MEMS SINS integrated navigation system as a background,take a research on integrated navigation filtering algorithm based on CKF.The main work is as follows:(1)As the existing ultra-tightly coupled integrated navigation system has a strong non-linearity and the large model error.So this paper studies the modeling of the ultra-tightly coupled integrated navigation system,and analyzes its observability.Because of the different observability of the carrier in different motion states,the rod arm error of the carrier is poorly observable,and in most cases it can be treated as a variable measurement noise,it can be lowered degree when going processing.So it can reduce the complexity of ultra-tightly coupled integrated navigation system model;(2)The ultra-tightly coupled integrated navigation system is complex,it contains both nonlinear variables,and linear variables,belonging to a conditional Gaussian filtering problem.Aiming at the characteristics of the ultra-tightly coupled integrated navigation system model,this paper firstly studies the conditional Gaussian filtering algorithm.The commonly used RBPF(Rao-Blackwellized Partical Filter)algorithm for solving the conditional Gaussian problem has a large computation complexity and prone to have particle degradation,which can lead to filtering divergence.In this paper,the conditional Gaussian filter based on 5CKF algorithm is studied.The algorithm uses the nonlinear variables in the 5CKF processing system,and the KF processing system uses the linear variables in the system to improve the accuracy of the algorithm while can reduce the computational complexity.Finally,the validity of the 5CKF/KF algorithm is verified in the integrated navigation filtering simulation environment.(3)When the carrier is in a time-varying noise environment,the noise model of the integrated navigation system can not be accurately described,which will lead the filter accuracy get down.Because the Interacting Multiple Model algorithm(IMM)is an effective method for time-varying noise problem,the IMM algorithm based on 5CKF is studied and analyzed in detail.Finally,the validity of the proposed IMM5 CKF algorithm is verified in the integrated navigation filtering simulation environment.(4)The vehicle experiments are carried out on the integrated GPS/MEMS SINS integrated navigation system,the collected data are filtered with 5CKF/KF algorithm and IMM5 CKF algorithm in the experiment and the experiment results with the proposed algorithm are compared with other nonlinear filters.The paper uses the vehicle experiment to verify the two algorithms can be effectively applied to the actual environment,and laid the foundation for further research.