A Real-Time Gravity Compensation Method for INS Based on BPNN
Duanyang Gao, Baiqing Hu, Fangjun Qin, Lubin Chang, Xu Lyu
Abstract
In order to solve the problem that the spherical harmonic model has a large amount of calculation, especially in the background of aviation navigation, it is difficult to realize real-time gravity compensation, a gravity compensation method based on back-propagation neural network (BPNN) is proposed. Firstly, the highest order/degree spherical harmonic model EIGEN-6C4 is used to calculate the gravity vector information in the planning navigation area, which is regarded as the true value. The gravity disturbance information and the corresponding position information are used as the training data set of BPNN. Then, the gravity model based on BPNN is obtained by training, and the model is applied to the gravity compensation of INS. Finally, the effectiveness of the method is verified by experiments. The numerical simulation results show that the standard deviations of east-west, north-south gravity disturbances errors and gravity anomaly errors are 0.29 mGal, 0.35 mGal and 0.81 mGal, respectively. The vehicle experiment results show that the horizontal radial position performance of INS after gravity compensation is improved by more than 26.68%, which is basically consistent with the gravity compensation using the highest order/degree spherical harmonic model, and ensures the real-time performance of the algorithm.