Hardware solution of the problem of stochastic robust discrete filtration in on-board locomotive control systems

   Today, the assessment of discrete nonlinear stochastic systems in the vast majority of cases is carried out on the basis of discrete stochastic filtering methods, which provide an optimal assessment of the measured state vector according to the root-mean-square criterion that is based on the discrete Kalman filter circuit and its various modifications.

   The main disadvantage of these filters is the need for an accurate a priori description of the probabilistic characteristics of the measurement interference of the estimated signal.

   At the same time, in real information-measuring and control systems on board locomotives operating under conditions of various disturbances, the statistical parameters of measurement noise either change randomly over time or are known approximately. In such a situation, the use of Kalman filtering methods is not possible. Due to the relevance of solving the filtering problem in a similar formulation, in this article, for discrete nonlinear stochastic systems perturbed by noise with unknown distribution densities, the problem of their robust recurrent estimation is solved based on a locally optimal criterion for the robustness of the estimate.

   An important feature of the developed robust estimation algorithm is its dimension, which coincides with the dimension of the object being assessed (while the dimension of modern filtering algorithms significantly exceeds the dimension of the object’s state vector).

   This makes it possible to dramatically reduce computational costs when implementing the proposed algorithm, which is very important for on-board navigation and control systems of locomotives.

1. Tikhonov, V. I. Statistical analysis and synthesis of radio engineering devices and systems / V. I. Tikhonov, V. N. Kharisov // Radio and communications. – 2004. – 304 p. – ISBN 5-256-01701-2.

2. Stochastic processing of satellite measurements in on-board navigation systems of locomotives / S. V. Sokolov, A. L. Okhotnikov, A. V. Kostyukov, O. I. Sokolova // Vestnik Rostovskogo Gosudarstvennogo Universiteta Putey Soobshcheniya. – 2022. – No. 2 (86). – P. 48–54. – DOI: 10.46973/0201-727X_2022_2_48.

3. The Kalman Filter Uncertainty Concept in the Possibility Domain, IEEE Trans. Instrum. Meas / A. Ferrero, R. Ferrero, W. Jiang, S. Salicone. – 68 (2019). – P. 4335–4347. – DOI: 10.1109/TIM.2018.2890317.

4. Structural identification of stochastic objects in intelligent dynamic systems / P. A. Kucherenko, L. N. Stazharova, A. V. Kostyukov [et al.] // Vestnik Rostovskogo Gosudarstvennogo Universiteta Putey Soobshcheniya. – 2022. – No. 3 (87). – P. 126–135. – DOI: 10.46973/0201-727X_2022_3_126.

5. Koshaev, D. A. Multi-alternative method for detecting and assessing violations based on the extended Kalman filter / D. A. Koshaev // Automation and Telemechanics. – 2010. – No. 5. – P. 70–83. – ISSN 0005-2310.

6. Patent No. 2755499 C1. Russian Federation, IPC H03H 21/00, G06E 3/00. Adaptive filtering method : No. 2021101912 : application. 01/28/2021 : publ. 09.16.2021 / S. V. Sokolov, A. A. Manin ; applicant for the Order of the Red Banner of Labor, Federal State Budgetary Educational Institution of Higher Education "Moscow Technical University of Communications and Informatics". – EDN: DWVVGV.

7. Copyright certificate No. 1800588 A1 USSR, IPC H03H 21/00. Adaptive Kalman filter: No. 4890558: app. 12/11/1990 : publ. 03/07/1993 / V. A. Alferenko, E. V. Borisov, N. G. Shebanov ; applicant MILITARY ACADEMY NAMED AFTER F. E. DZERZHINSKY. – EDN: SIEFYS.

8. Wang, D. Augmented Cubature Kalman filter for nonlinear RTK/MIMU integrated navigation with non-additive noise / D. Wang, H. Ly, J. Wu. – Measurement 97, 2017. – P. 111–125.

9. Särkkä, S. Recursive noise adaptive Kalman filtering by variational Bayesian approximations / S. Särkkä, A. Nummenmaa // IEEE Transactions on Automatic control. – 2009. – Vol. 54. – P. 596–600. – DOI: 10.1109/tac.2008.2008348.

10. Kalman filtering based on the maximum correntropy criterion in the presence of non – Gaussian noise / R. Izanloo, S. A. Fakoorian, H. S. Yazdi, D. Simon // Annual Conference on Information Science and Systems (CISS), Princeton, USA : Proccedings. – 2016. – P. 500–505. – DOI: 10.1109/ciss.2016.7460553.

11. Gao, H. Induced l2 and generalized H∞ filtering for systems with repeated scalar nonlinearities / H. Gao, J. Lam, C. Wang // IEEE Transact. Signal Proc. – 2005. – V. 53. – No. 11. – P. 4215–4226. – DOI: 10.1109/TSP.2005.857009.

12. Gao, H. New approach to mixed H2/H∞ filtering for polytopic discrete-time systems / H. Gao, J. Lam, C. Wang // IEEE Transact. Signal Proc. – 2005. – V. 53. – No. 8. – P. 3183–3192. – DOI: 10.1109/TSP.2005.851116.

13. Gao, H. Robust H∞ filtering for discrete stochastic time-delay systems with nonlinear disturbances / H. Gao, J. Lam, C. Wang // Nonlinear Dynam. Syst. Theory. – 2004. – V. 4. – No. 3. – P. 285–301. – DOI: 10.1080/03081079.2014.973730.

14. Poor, H. V. Minimax State Estimation for Linear Stochastic Systems with Noise Uncertainty / H. V. Poor, D. P. Looze // IEEE Trans. Automat. Control. – 1981. – V. 26. – P. 902–906. – DOI: 10.1109/TAC.1981.1102756.

15. Sayed, A. H. A Framework for State-space Estimation with Uncertain Models // IEEE Trans. Automat. Control. – 2001. – V. 46. – P. 998–1013. – DOI: 10.1109/9.935054.

16. Calafiore, G. Minimum Variance Estimation with Uncertain Statistical Model / G. Calafiore, L. El Ghaoui // Proc. IEEE CDC. – 2001. – P. 3497–3499. – DOI: 10.1109/CDC.2001.980400.

17. Miller, B. M. Robust estimation based on the method of least moduli and the Kalman filter / B. M. Miller, K. S. Kolosov // Automation and Telemechanics. – 2020. – No. 11. – P. 72–92. – DOI: 10.31857/S0005231020110057.

18. Patent No. 2804256 C1 Russian Federation, IPC G06F 17/16, H03H 19/00. Robust discrete stochastic filter : No. 2023101070 : application. 01. 19. 2023 : publ. 09. 26. 2023 / S. V. Sokolov, I. V. Reshetnikova ; applicant "Moscow Technical University of Communications and Informatics". – EDN: CQNCWZ.

19. Smirnov, Yu. A. Fundamentals of microelectronics and microprocessor technology : textbook / Yu. A. Smirnov, S. V. Sokolov, E. V. Titov. – 2^nd, Corrected. – Saint Petersburg : Lan Publishing House, 2013. – 496 p. – ISBN 978-5-8114-1379-9.

20. Patent No. 2022328 C1 Russian Federation, IPC G06E 1/04. Optical multiplier : No. 5045839/24 : appl. 06/03/1992 : publ. 10.30.1994 / S. V. Sokolov ; applicant Rostov Higher Military Command and Engineering School of Missile Forces named after. Chief Marshal of Artillery M. I. Nedelin – EDN: WBQHTV.