The integrated navigation systems accuracy improving by machine learning techniques Kalman filter parameters adaptive tuning

Аuthors

Pron'kin A. N.^*, Zharkov M. V.^**, Kuznetsov I. M.^***

Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia

*e-mail: pronkinan@mai.ru
**e-mail: mv_zharkov@mai.ru
***e-mail: kuznetsovim@mai.ru

Abstract

The article addresses the problem of improving the accuracy of integrated inertial-satellite navigation systems operating under conditions of incomplete a priori information about the measurement noise statistics. An adaptive filtering algorithm based on the analysis of the Kalman filter innovation sequence and the application of gradient boosting algorithms (LightGBM) is proposed. Using the statistical consistency criterion, an approach to forming a feature space for machine learning is formalized. Experimental studies based on a microelectromechanical (MEMS) inertial measurement unit have shown the high efficiency of the trained model in predicting the intensities of sensor noise, which significantly reduces the errors in estimating navigation parameters.

Keywords:

inertial navigation system; Kalman filter; adaptive filtering; machine learning; noise covariance matrix; innovation sequence

References

1. Influence of Longitudinal Acceleration Parameters on the Ability of Instrument Error Estimation of Inertial Navigation System / K. K. Veremeenko, M. V. Zharkov, I. M. Kuznetsov, A. N. Pron'kin // Russian Aeronautics. – 2023. – Vol. 66, No. 1. – pp. 115-124. – DOI 10.3103/s1068799823010166. – EDN ULQZLT.
2. I. M. Kuznetsov et al  Moving object SINS transfer alignment time synchronization parameters estimation, 2021, J. Phys.: Conf. Ser. 1925 012025, https://doi.org/ 10.1088/1742-6596/1925/1/012025.
3. G. I. Yemelyantsev, A. P. Stepanov «Integrated inertial-satellite orientation and navigation systems» St. Petersburg: Concern Central Scientific Research Institute Electropribor", 2016. – pp 393. ISBN 978-5-91995-029-5. 
4. О. А. Stepanov «Fundamentals of estimation theory with applications to navigation information processing tasks» Vol. 1. Edition 3. – St. Petersburg: Concern Central Scientific Research Institute Electropribor ", 2017. – p 509. – ISBN 978-5-91995-051-6. – EDN YNTWDR.
5. О. А. Stepanov «Fundamentals of estimation theory with applications to navigation information processing tasks» Vol. 2. Edition 3. – St. Petersburg: Concern Central Scientific Research Institute Electropribor ", 2017. – p 428. – ISBN 978-5-91995-053-0. – EDN YNUDBJ.
6. Investigation of Parametric Uncertainty Influence on Accuracy of Strapdown Inertial Navigation System Transfer Alignment / K. K. Veremeenko, M. V. Zharkov, I. M. Kuznetsov, A. N. Pron'kin // Russian Aeronautics. – 2021. – Vol. 64, No. 3. – pp. 518-525. – DOI 10.3103/S106879982103020X. – EDN AQGYGH. 
7. Kuzovkov N.T., Karabanov S.V., Salychev O.S. «Continuous and discrete control systems and identification methods», Moscow, 1978. p. 222. 
8. Salychev O.S. Algorithm for preventing divergence of the optimal filtration process. – Technical cybernetics, 1979, Vol.4, pp. 168-170. 
9. Quigley A. An approach to the control of divergence in Kalman Filter algorithms. – Int. J. Control, 1973, vol. 17, p. 741-746.
10. Mehra R. Kalman identification and adaptive filtering – Mechanics Vol 3, 1971, pp. 34-51.
11. Jazwinski A.H. Stochastic Processes and Filtering Theory, - N.Y. 1970, p. 376.
12. Chernodarov A.V., Ivanov S.A. «Identification of models and adaptive filtering of inertial sensors noises», Trudy MAI, Vol. 99, 2018.
13. H. Wang et al., "Dynamic Process Noise Covariance Adjustment in GNSS/INS Integrated Navigation Using GRU-SAC for Enhanced Positioning Accuracy," in IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 6, pp. 9047-9064, June 2025, doi: 10.1109/TITS.2025.3540205, 
14. Z. Zou, T. Huang, L. Ye and K. Song, "CNN Based Adaptive Kalman Filter in High-Dynamic Condition for Low-Cost Navigation System on Highspeed UAV," 2020 5th Asia-Pacific Conference on Intelligent Robot Systems (ACIRS), Singapore, 2020, pp. 103-108, https://doi.org/10.1109/ACIRS49895.2020.9162601, 
15. Or, Barak & Klein, Itzik. (2022). A Hybrid Model and Learning-Based Adaptive Navigation Filter. https://doi.org/10.48550/arXiv.2207.12082
16. Zhang L., Sidoti D., Bienkowski A., Pattipati K. R., Bar-Shalom Y., Kleinman D. L. On the Identification of Noise Covariances and Adaptive Kalman Filtering: A New Look at a 50 Year-old Problem//IEEE Access. – 2020. – Vol. 8.
17. Fang W., Jiang J., Lu S. et al. A LSTM Algorithm Estimating Pseudo Measurements for Aiding INS during GNSS Signal Outages//MDPI Remote Sensing. – 2020. – Vol. 12, No. 2.
18. Assad, A. Adaptation of Noise Covariance in Extended Kalman Filter Using Reinforcement Learning for Improved UAV Attitude Estimation / A. Assad, S. A. Serikov // Gyroscopy and Navigation. – 2025. – Vol. 16, No. 3. – P. 228-239. – DOI 10.1134/S2075108725700269. – EDN UVRLTK.
19. Navigation systems of unmanned aerial vehicles with artificial intelligence units / K. K. Veremeenko, M. V. Zharkov, I. M. Kuznetsov, A. N. Pron'kin, R.Yu. Zimin //XXXI St. Petersburg International Conference on Integrated Navigation Systems: Collection of reports, St. Petersburg: Concern Central Scientific Research Institute Electropribor", 2024. – pp. 101-105. 
20. Navigation system neural network architecture parameters preformance influence investigation / I. M. Kuznetsov, K. K. Veremeenko, M. V. Zharkov, A. N. Pron'kin, R.Yu. Zimin //XXXII St. Petersburg International Conference on Integrated Navigation Systems: Collection of reports, St. Petersburg: Concern Central Scientific Research Institute Electropribor", 2025. – pp. 56-62.
21. Rivkin S.S., Ivanovsky R.I., Kostrov A.V. Statistical optimization of navigation systems. L.: Shipbuilding, 1976. p.280.
22. Novoselova, O. E., R. A. Zavyalov, I. V. Muratov «Investigation of adaptive algorithms of recurrent filtering in processing data from GNSS in case of loss of radio signals from navigation satellites» // Automation. Modern technologies. – 2024. – Vol. 78, Edition 1. – pp. 23-29. – DOI 10.36652/0869-4931-2024-78-1-23-29.
23. M. Zharkov, K. Veremeenko, I. Kuznetsov, A. Pronkin «Experimental Results of Attitude Determination Functional Algorithms Implementation in Strapdown Inertial Navigation System»//Sensors 2022, 22, 1849
24. Laboratory for Research and Development of Navigation and Orientation Systems. access: https://mai.ru/science/studies/equipment/index.php?ELEMENT_ID= 167952.
25. NovAtel Inertial Explorer. access: https://novatel.com/products/waypoint-post-processing-software/inertial-explorer.

Download