Research On Inertial/GNSS/Polarized Light Compass Combined Navigation Method

In recent years,the importance of unmanned mobile platform in civil,military and other fields is increasing,and the research of unmanned platform has gradually become one of the hot spots of competition among countries.Navigation is a very important part of unmanned platform,so the stability and accuracy of the unmanned navigation system plays a decisive role in unmanned platform.Polarized light is a new method of navigation.The navigation information he USES comes from nature,has the characteristic of stable distribution,cannot be destroyed artificially in a short time,and can realize stable autonomous navigation.However,the development of polarized light navigation still has some limitations,so in order to obtain a stable and highprecision polarized light navigation system,this paper proposes a mems-ins/polarized light /GNSS seamless integrated navigation system,which mainly includes three parts: seamless navigation model construction,information fusion algorithm,and seamless anti-interference modeling.1)construction of navigation model: a multi-sensor integrated navigation system model was proposed and a multi-sensor information receiving and processing platform with TX2 as the core was built,which was loaded on a homemade unmanned vehicle for testing.2)information fusion algorithm: a new adaptive Kalman filter for nonlinear systems is proposed.The adaptive filter estimates the covariance of the system,improves the robustness,precision,frequency and reduces the computation.The experimental results show that the model has the same high and low frequency navigation accuracy,and ensures the high frequency output of the system.Compared with other algorithms,the mean output error of the system increased from 4.81 to 2.18,which decreased by 54.7%.The root-mean-square error was reduced from 2.83 to 1.41,an increase of 50.2%.3)seamless anti-interference modeling: a seamless navigation dual model based on square root cubic Kalman filter(SRCKF)and random forest regression(RFR)was proposed.Using this method,the seamless navigation ability of the system can be guaranteed even when the measuring signal is interrupted.In the proposed dual model,a sub-model 1 which directly relates the specific force of the inertial navigation system to the measurement result of the filter and a sub-model 2 which directly relates the volume point and new information of SRCKF to the error caused by the filter are established.Combined with the SRCKF and RFR algorithms,the dual-model system can seamlessly predict and estimate the vehicle's speed and position when the measurement signal is blocked.The experimental results show that the navigation accuracy of the proposed model is 40% higher than that of the traditional algorithm.