Research For Geomagnetic Navigation Method Based On Magnetometer Difference Measurements And Model Reconstruction & Compensation

Geomagnetic navigation plays an important role in the autonomous navigation of near-Earth satellites,because it has the following advantages: strong autonomy,good concealment,low cost,ready-to-use,and strong anti-jamming ability,not accumulating navigation errors over time.In this paper,a geomagnetic navigation method of near-Earth satellites based on measurement differencing and model reconstruction and compensation is studied.The innovative results obtained are as follows:First of all,a geomagnetic navigation algorithm based on a measurement differencing extended Kalman filter(MDEKF)is proposed to address the problem that geomagnetic storms can change measurements of geomagnetic field,which leads to a decrease in the accuracy and stability of geomagnetic navigation.The MDEKF algorithm can effectively reduce unknown measurement errors caused by geomagnetic storms in the geomagnetic navigation of near-Earth satellites,thereby maintaining the accuracy and stability of the navigation algorithm.The performance of the MDEKF algorithm and the EKF algorithm under different levels of geomagnetic storms and different sampling intervals are compared with real flight data from Swarm A.The simulation results show that as the level of the geomagnetic storm increases,the advantages of the MDEKF algorithm over the EKF algorithm are more obvious.Specifically,when measurements are geomagnetic field vectors and the sampling interval is relatively small,the maximum position estimation error obtained by the MDEKF algorithm is half of the result by the EKF algorithm.Secondly,a geomagnetic navigation algorithm based on a rapid reconstruction of a three-dimensional decoupled geomagnetic field(TDGF)model is propose,aiming at the shortcomings of the IGRF model,such as a long update interval,decreasing accuracy over time,unable to respond to the pole shift quickly.The TDGF model can be updated quickly and effectively eliminate time accumulation errors of the IGRF model.On the one hand,simulations are performed using real flight data of Swarm B and Swarm C.The results show that when the TDGF model update interval is two months,the performance of the geomagnetic navigation algorithm based on the TDGF model is equivalent to that based on the initial IGRF model.Specifically,when measurements are geomagnetic vectors,the navigation accuracy is about 1km,and when geomagnetic magnitudes,the navigation accuracy is about 3km.On the other hand,simulation data are used to study the accuracy potential of the algorithm when the algorithm is applied to near-Earth micro-nano satellites.Finally,a geomagnetic navigation algorithm based on a neural network geomagnetic field error compensation(NNGFEC)model is proposed to overcome the systematic error of a spherical harmonic model describing actual geomagnetic field.By a neural network,the algorithm transforms the geomagnetic latitude information into the corresponding compensation value of geomagnetic model error.In other words,the new mathematical model is used to compensate the spherical harmonic function model,thus improving the accuracy of the geomagnetic field model.Using real flight data of Swarm B and Swarm C,the geomagnetic navigation algorithm based on the NNGFEC model is simulated and verified.The results show that: in terms of stability,convergence speed and accuracy,the faster NNGFEC model is updated,the more significant the advantage of the geomagnetic navigation algorithm based on the compensation model will be.In addition,when the model updating interval is one month,the RMS of position estimation errors obtained by the compensation model is only half of that by the IGRF model.