LSTM-Based Shuler Oscillation Error Restriction for Shipborne SINS

The Shuler oscillation error generated by SINS in pure inertial solution seriously affects the accuracy of navigation parameters such as attitude, velocity, and position, which is very unfavorable for long-term SINS. The conventional external/internal damping technology and integrated navigation technique are able to limit SINS Schuler oscillation error, but destroy SINS’ autonomy and overly rely on the accuracy of external auxiliary information. According to the principle that two periodic signals with phase difference $$\pi$$ can be cancelled on average, two sets of navigation algorithms are established by using the method of deep learning to eliminate the Schuler oscillation error while maintaining the autonomy of the SINS. The simulation verification indicates that the proposed algorithm has a good restriction effect for shipborne SINS Schuler oscillation error.