Position Calibration Technology for Long-Term Endurance Inertial Navigation Systems Based on Sparse and Low-Precision Position Information

To enhance the long-term endurance navigation and positioning capabilities of inertial navigation systems (INSs), external auxiliary information is typically required for calibration. However, in certain challenging environments, obtaining continuous reliable and high-precision auxiliary information becomes difficult, and only sparse and low-precision position information may be available. Investigating methods to utilize this unfavorable information for calibrating INSs holds significant practical value. Building upon the inherent error characteristics of INSs, this study introduces a novel position calibration methodology. The first step is to make a comprehensive judgment by combining the position calculated by the INS and the reference position and selecting the usable sparse and low-precision reference position that meets the judgment conditions. The second step is to calibrate the INS with reference positions that meet the judgment conditions. Finally, a simulation static test and dynamic test were carried out; in the simulation static test, the std value of the latitude error decreased from 0.3250 n mile to 0.2245 n mile, and the accuracy improved by 30.92%; the max value of the longitude error decreased from 8.2168 n mile to 1.1629 n mile, and the accuracy improved by 85.85%; and the max value of the position error decreased from 8.2168 n mile to 1.2003 n mile, and the accuracy improved by 85.39%. In the simulation dynamic test, the std value of the latitude error decreased from 0.4501 n mile to 0.2384 n mile, and the accuracy improved by 47.03%; the max value of the longitude error decreased from 3.5907 n mile to 0.9383 n mile, and the accuracy improved by 73.87%; and the max value of the position error decreased from 3.5931 n mile to 1.0581 n mile, and the accuracy improved by 70.55%. The results of these tests show that the cumulative error which grows with time in INSs is obviously suppressed.