Research On SINS/CNS Deeply Integrated Navigation System And High-performance Navigation Filtering Method

Navigation technology is an important supporting technology for weapon equipment informatization.With the increasing demand for information warfare,the need for high-precision,high-reliability,full-autonomous,all-day navigation is becoming more and more urgent.As a branch of navigation technology,SINS/CNS integrated navigation system plays an indispensable role in the aerospace field of medium and long-range missiles,aircrafts and satellites and so on due to its inherent advantages in accuracy,reliability and autonomy.SINS/CNS integrated navigation is a multi-sensor data fusion problem with nonlinearity and uncertainties and other factors.It is an effective way to realize the highperformance navigation based on studying the integrated navigation system with multimeasurement equipment and the high-performance navigation method suitable for dealing with nonlinearity and uncertainties.In this paper,the missile-loading platform is used as the navigation object.For the SINS/CNS integrated navigation system,the methods of improving its navigation performance are systematically studied from the data layer,model layer,algorithm layer and application layer.The main work is as follows:On the data layer,for the sensor measurement outlier problem existing in the actual integrated navigation system,the outlier processing method of a variety of different measurement data is comprehensively analyzed from the aspects of the outlier types,the algorithm complexity,and the outlier eliminating efficiency.The optimal navigation data outlier processing method is given,which is the multi-point linear forecasting method based on adaptive threshold.On the model layer,the SINS/CNS deep integrated navigation scheme is designed for compensating accelerometer accumulation error of the traditional SINS/CNS integrated navigation system,using two star sensors and an altimeter to add new measurement information which enables the navigation system to satisfy the system observability requirements and improve the positional accuracy of the navigation target.For the inconsistency between the gyroscope and the star sensor installation benchmark,they are theoretically analyzed in terms of different benchmarks and non-integration.The error influencing relationship in the mode is analyzed in theory and the best star sensor-gyroscope with the same reference integrated installation structure is given.On the algorithm layer,according to the nonlinearity,uncertainty and non-Gaussian of the system noise,the corresponding navigation filtering algorithms based on model error and measurement noise suppression are designed.For the measure outliers and the non-Gaussian noises of the measurement,the unscented Kalman filter algorithm based on the maximum correlation entropy criterion is used to deal with the uncertainty of the measurement model,and the influence function is used to prove the robustness of the filtering algorithm in theory,indicating that it can effectively suppress the influence of nonGaussian noise and outliers of the measurent on navigation performance.For the state perturbation problem existing in navigation system,a strong tracking unscented Kalman filter algorithm is used to adjust the state estimation covariance matrix in real-time.Simulations indicate that it can effectively suppress the problem of state estimation error divergence and acceleration cumulative deviation caused by state perturbation.On the application layer,combined with the SINS/CNS deep integrated navigation system and high-performance navigation filtering algorithm studied above,they are applied to the navigation problem in the case of multi-model state correlation in the SINS/CNS deeply integrated navigation system.The interacting multiple model Kalman filter is improved based on Kullback-Leibler divergence to achieve tracking and positioning navigation targets.And the adaptive fading factor Kalman filter and the maximum correntropy Kalman filter are fused in multiple model interaction.It effectively improves the navigation positioning accuracy under the condition of simultaneous state model uncertainty and measurement model uncertainty and provides the theoretical and technical support for the practical application of SINS/CNS integrated autonomous navigation.