An Integrated Navigation Algorithm Based on LSTM Neural Network

For the issue that the positioning accuracy of GNSS (Global Navigation Satellite System)/INS (Inertial Navigation System) integrated navigation system is reduced because of the presence of Satellite blocking area in complex environment, a method based on LSTM (Long-Short Term Memory) neural network is designed to improve system. While satellite valid, the GNSS position increments are treated as the true value to train, and when the satellites reject, the GNSS information is replaced by the accumulated position increment predicted by the trained network to correct the accumulated error of INS. Experiments are based on the emulational IMU information, and the consequence show that the proposed algorithm can effectively suppress the divergence of pure INS caused by the loss of satellite signals, and improve the accuracy of the system.