Research On SINS/USBL Integrated Navigation Technology Based On Adaptive Student-t Kalman Filter

After decades of development,the ultra-short baseline underwater acoustic positioning system has been widely used in underwater vehicles.In order to further improve the navigation and positioning accuracy of underwater navigation,a large number of equipment adopts integrated navigation methods in practice.Strapdown inertial navigation system and ultra-short baseline hydroacoustic positioning system for combined navigation can obtain better underwater navigation accuracy.However,due to the limitation of the propagation speed of the underwater acoustic signal,the measurement delay will be caused,and the propagation multipath will lead to the generation of measurement outliers,which will lead to non-stationary measurement noise,which greatly contributes to the positioning accuracy and stability of the integrated navigation system.Aiming at the above-mentioned problems,this paper conducts research on the underwater SINS/USBL integrated navigation system.The specific research contents are as follows:First,the positioning principle of the strapdown inertial navigation system and the ultra-short baseline system is explained in detail,and the corresponding error model is given.The specific expression of the SINS/USBL integrated navigation system model is established through the error equation,and the Kalman filter is used to perform Simulations under Gaussian noise environment and measurement outlier environment.The simulation results show that under Gaussian noise conditions,Kalman filtering can get better estimation results,and the ultra-short baseline hydroacoustic positioning system has a better error in the strapdown inertial navigation system.A good auxiliary correction effect can prevent the error estimate from accumulating over time.In the presence of measurement outliers,the Kalman filter is used to simulate the SINS/USBL integrated navigation system with measurement outliers,and it is found that the measurement outliers have a huge interference to the Kalman filter,and the accuracy of the Kalman filter is very large.decline.Aiming at this problem,the measured noise is modeled as a Student-t distribution,and the VB method is used to infer,a SINS/USBL integrated navigation method based on Student-t Kalman filter is designed,and the integrated navigation system is simulated through simulation.Afterwards,for the problem of non-stationary measurement noise distribution under actual application environment changes,an adaptive Gaussian-Student-t Mixture Kalman Filter(GSAKF)SINS/USBL integrated navigation method was designed,Introduce Bernoulli distribution to describe the Gaussian-Student-t mixed measurement noise probability density distribution,and use the VB method to jointly infer the state vector,auxiliary parameters,and the posterior probability estimation of Bernoulli random variables.The simulation verifies the effectiveness of the method in non-stationary noise environment.Finally,in view of the signal delay of the ultra-short baseline hydroacoustic positioning system caused by the propagation delay of the underwater acoustic signal,an improved SINS/USBL integrated navigation method based on the Student-t Kalman filter was designed to combine the strapdown inertial navigation system with the ultra-short The position calculated by the baseline hydroacoustic positioning system is synchronized with the time.The simulation shows that the improved Kalman filter based on Student-t can improve the accuracy of the integrated navigation system under the condition of delay measurement.