DNN-Based AI-Driven H2/H∞ Filter Design of Nonlinear Stochastic Systems via Two-Coupled HJIEs-Supervised Adam Learning Algorithm

Authors

  • Bor-Sen Chen

    Department of Electrical Engineering, National Tsing Hua University, Hsinchu 300044, Taiwan
  • Jui-Ming Ma

    Department of Electrical Engineering, National Tsing Hua University, Hsinchu 300044, Taiwan
  • Ruei-Syuan Wu

    Department of Electrical Engineering, National Tsing Hua University, Hsinchu 300044, Taiwan

DOI:

https://doi.org/10.30564/aia.v6i1.7261
Received: 2 September 2024 | Revised: 15 September 2024 | Accepted: 20 September 2024 | Published Online: 20 October 2024

Abstract

This study introduces a new approach using supervised learning deep neural networks (DNNs) to develop an AI-driven filter for nonlinear stochastic signal systems with external disturbance and measurement noise. The filter aims to achieve a balanced design between and norm of the state estimation error to achieve both optimal and robust filtering design of nonlinear signal system simultaneously while considering environmental disturbance and measurement noise. Traditionally, this nonlinear  filter design involves solving complex two-coupled Hamilton-Jacobi-Issac Equations (HJIEs). To simplify this complicated design process, a novel two-coupled HJIEs-supervised Adam learning algorithm is proposed for DNN-based AI-driven filter. This algorithm trains a  DNN-based AI-driven filter offline using worst-case scenarios of environmental disturbance and measurement noise. This training phase generates state estimation errors that teach the DNN-based AI-driven filter how to coordinate nonlinear system model with worst-case external disturbance and measurement noise, Luenberger-type filter, estimation error dynamic model and two-coupled HJIEs-supervised deep Adam learning algorithm to achieve the mixed  filtering strategy effectively. The study demonstrates theoretically that this approach will achieve the desired mixed  filtering strategy once the Adam learning algorithm converges. Finally, the effectiveness of the proposed DNN-based AI-driven filter design method is validated through simulations, specifically involving trajectory estimation and prediction of an incoming ballistic missile detected by a radar system.

Keywords:

DNN-based AI filter; Deep neural network (DNN); Mixed H2/H∞ filter; Hamilton-Jacobi-Isaacs Equation (HJIE); Nonlinear stochastic system and HJIE-supervised adam learning algorithm

References

[1] Anderson, B.D.O., Moore, J.B., 1979. Optimal Filtering. Prentice Hall: Englewood Cliffs, NJ, USA.

[2] Khargonekar P.P., Rotea M.A., Baegens E., 1996. Mixed filtering. International Journal of Robust and Nonlinear Control. 6(6), 313–330.

[3] Gao H., Lam J., Xie L., et al., 2005. New approach to mixed filtering for polytopic discrete-time systems. IEEE Transactions on Signal Processing. 53(8), 3183–3192. DOI: https://doi.org/10.1109/TSP.2005.851116

[4] Li, H., Fu, M., 1997. A linear matrix inequality approach to robust filtering. IEEE Transactions on Signal Processing. 45(9), 2338–2350. DOI: https://doi.org/10.1109/78.622956

[5] Palhares, R.M., Peres, P.L.D., 2001. LMI approach to the mixed filtering design for discrete-time systems. IEEE Transactions on Aerospace and Electronic Systems. 37(1), 292–296. DOI: https://doi.org/10.1109/7.913689

[6] Zhang, W., Chen, B.S., Tseng, C.S., 2005. Robust filtering for nonlinear stochastic systems. IEEE Transactions on Signal Processing. 53(2), 589–599. DOI: https://doi.org/10.1109/WCICA.2008.4593266

[7] Chen, B.S., Tsai, C.L., Chen, D.S., 2003. Robust and mixed filter for equalization designs of nonlinear communication systems: Fuzzy interpolation approach. Fuzzy interpolation approach. IEEE Transactions on Fuzzy Systems. 11(3), 384–398. DOI: https://doi.org/10.1109/TFUZZ.2003.812698

[8] Chen, B.S., Chen, W.H., Wu, H.L., 2009. Robust mixed global linearization filter design for nonlinear stochastic systems. IEEE Transactions on Circuits and Systems I. 56(7), 1441–1454. DOI: https://doi.org/10.1109/TCSI.2008.2007059

[9] Chen, B.S., Wu, C.F., 2015. Robust scheduling filter design for a class of nonlinear stochastic system. IEEE Transactions on Signal Processing. 63(23), 6245–6257. DOI: https://doi.org/10.1109/TSP.2015.2465305

[10] Chen, B.S., Lee, H.C., Wu, C.F., 2015. Pareto optimal filter design for nonlinear stochastic systems via multiobjective optimization. IEEE Transactions on Fuzzy Systems. 23(2), 387–399. DOI: https://doi.org/10.1109/TFUZZ.2014.2312985

[11] Dahl, G.E., Yu, D., Dang, L., Acero, A., 2012. Contex-dependent pretrained deep neural networks for large-vocabulary speech recognition. IEEE Transactions on Audio, Speech, and Language Processing. 20(1), 30–42. DOI: https://doi.org/10.1109/TASL.2011.2134090

[12] Eshratifar, A.E., Esmaili, A., Pedram, M., 2019. BottleNet: A deep learning architecture for intelligent mobile cloud computing service. ISLPED 2019, Lausanne, Switzerland. pp. 1–6.

[13] Maas, A.L., Qi, P., Xie, Z., et al., 2017. Building DNN acoustic models for large vocabulary speech recognition. Computer Speech and Language. 41, 195–213. DOI: https://doi.org/10.1016/j.csl.2016.06.007

[14] Shin H.-C., Roth, H.-R., Gao, M., et al, 2016. Deep convolutional neural networks for computer aided detection: CNN architectures dataset characteristics and transfer learning. IEEE Transactions on Medical Imaging. 35(5), 1285–1298. DOI: https://doi.org/10.1109/TMI.2016.2528162

[15] Sutton, R.S., Barto, A.G., 2015. Reinforcement Learning: An Introduction, 2nd ed., in progress, Bradford Books.

[16] Wooldridge M., 2022. What is missing from contemporary AI? The world Intelligent Computing. 2022, 9847630.

[17] Yin, J., Su, S., Xun, J., et al., 2020. Data-driven approaches for modeling train control: Comparison and case studies. ISA Transactions. 98(5), 349–363. DOI: https://doi.org/10.1016/j.isatra.2019.08.024

[18] Bertsekas Dimitri, P., 2019. Reinforcement Learning and Optimal Control. Athena Scientific.

[19] Chen, B.S., Wu, P.S., Lee, M.Y., 2021. Robust deep neural network based filter design of nonlinear stochastic systems. IEEE Access. 9, 165103–1651199. DOI: https://doi.org/10.1109/ACCESS.2021.3133899

[20] Zhang, W., Xie, L., Chen, B.S., 2015. Stochastic Control. CRC Press.

[21] Sakamoto, S., Vander Schaft, A.J., 2008. Analytical approximation methods for the stabilization solution of Hamilton-Jacobi equation. IEEE Transactions on Automatic Control. 53(10), 2335–2350. DOI: https://doi.org/10.1109/TAC.2008.2006113

[22] Dauphin, Y.N., Vries, H.D., Chung, J., et al., 2015. RMSProp and equilibrated adaptive learning rates for non-convex optimization. Neural Networks. arxiv preprint. arxiv:1502.04390. pp. 1–9.

[23] Kingma, D.P., Ba, J., 2014. Adam: A method for stochastic optimization. ICLR. San Diego. arxiv preprint. arxiv:1412.6980. pp. 1–15.

[24] Cui, H., Liu, W., Zhu, S., et al., 2016. A hypersonic vehicle tracking algorithm based on the UKF Generalized labeled multi bernoulli filter. Proceedings of the 2016 35th Chinese Control Conference (CCC); Chengdu, China; 27–29 July 2016.. Volume 35, pp. 4911–4916.

[25] Siouris, G.M., Chen, G., Wang, J., 1997. Tracking an incoming ballistic missile using and extended internal kalman filter. IEEE Transactions on Aerospace and Electronic Systems. 33(1), 232–240. DOI: https://doi.org/10.1109/7.570753

[26] Chen, B.-S., Lee, M.-Y., Chen X.-H., 2020. Security-enhanced filter design for stochastic systems under malicious attack via smoothed signal model and multiobjective estimation method. IEEE Transactions on Signal Processing. 68, 4971–4986. DOI: https://doi.org/10.1109/TSP.2020.3019136

Downloads

How to Cite

Chen, B.-S., Ma, J.-M., & Wu, R.-S. (2024). DNN-Based AI-Driven H2/H∞ Filter Design of Nonlinear Stochastic Systems via Two-Coupled HJIEs-Supervised Adam Learning Algorithm. Artificial Intelligence Advances, 6(1), 34–55. https://doi.org/10.30564/aia.v6i1.7261

Issue

Vol. 6 , Iss. 1 (October 2024)

Article Type

Article

License

Copyright © 2024 Bor-Sen Chen, Jui-Ming Ma, Ruei-Syuan Wu

Creative Commons License

This is an open access article under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) License.

Crossref
Scopus
Google Scholar
Europe PMC