Journal of the Korea Institute of Military Science and Technology (한국군사과학기술학회지)

The Korea Institute of Military Science and Technology (한국군사과학기술학회)
권호 표지
DOI QR코드

DOI QR Code

Application of Recurrent Neural-Network based Kalman Filter for Uncertain Target Models

불확정 표적 모델에 대한 순환 신경망 기반 칼만 필터 설계

  • DongBeom Kim (Department of Electrical Engineering, Hanyang University) ;
  • Daekyo Jeong (Missile Research Division, Agency for Defense Development (ADD)) ;
  • Jaehyuk Lim (Missile Research Division, Agency for Defense Development (ADD)) ;
  • Sawon Min (Land Radar Team, Hanwha Systems Co., Ltd.) ;
  • Jun Moon (Department of Electrical Engineering, Hanyang University)
  • 김동범 (한양대학교 전기공학과) ;
  • 정대교 (국방과학연구소 미사일연구원) ;
  • 임재혁 (국방과학연구소 미사일연구원) ;
  • 민사원 (한화시스템(주) 지상레이다팀) ;
  • 문준 (한양대학교 전기공학과)
  • Received : 2022.11.17
  • Accepted : 2023.01.30
  • Published : 2023.02.05
Download PDF
⟨ Previous Next ⟩

Abstract

For various target tracking applications, it is well known that the Kalman filter is the optimal estimator(in the minimum mean-square sense) to predict and estimate the state(position and/or velocity) of linear dynamical systems driven by Gaussian stochastic noise. In the case of nonlinear systems, Extended Kalman filter(EKF) and/or Unscented Kalman filter(UKF) are widely used, which can be viewed as approximations of the(linear) Kalman filter in the sense of the conditional expectation. However, to implement EKF and UKF, the exact dynamical model information and the statistical information of noise are still required. In this paper, we propose the recurrent neural-network based Kalman filter, where its Kalman gain is obtained via the proposed GRU-LSTM based neural-network framework that does not need the precise model information as well as the noise covariance information. By the proposed neural-network based Kalman filter, the state estimation performance is enhanced in terms of the tracking error, which is verified through various linear and nonlinear tracking problems with incomplete model and statistical covariance information.

Keywords

Acknowledgement

이 논문은 2022년도 한화시스템(주)의 재원을 지원 받아 수행된 연구임. 본 논문 발전을 위해 심사를 해주신 심사위원님들께 감사말씀을 드립니다.

References

  1. H. Kang, "Direction of Operation of ROKA RPA for Future Warfare," Korea Institute for Defense Analyses, Vol. 35, No. 1, pp. 7-33, 2019.
  2. Market Forecast, "Military Robots and Autonomous Systems - Market and Technology Forecast to 2030," 2020.
  3. Y. Heo, D. J. Kim, I. R. Lee, D. W. Kim, and Y. S. Jung, "R&D Trends and Technology Development Plan on Military Drone," Journal of the Korea Academia-Industrial Cooperation Society, Vol. 22, No. 12, pp. 324-331, 2021. https://doi.org/10.5762/KAIS.2021.22.12.324
  4. M. Mallick, V. Krishnamurthy, and B.-N. Vo, Integrated Tracking, Classification, and Sensor Management, IEEE Press, 2013.
  5. D. Simon, "Optimal State Estimation," Wiley, 2006.
  6. J. Moon, "Design of the Target Estimation Filter based on Particle Filter Algorithm for the Multi-Function Radar," Journal of the Korea Institute of Military Science and Technology, Vol. 14, No. 3, pp. 517-523, 2011. https://doi.org/10.9766/KIMST.2011.14.3.517
  7. B. S. Kang and I. S. Lee, "Estimation of a Projectile Impact Point Using Radar Measurement and Extended Kalman Filte," Journal of Korean Institute of Information Technology, Vol. 18, No. 2, pp. 101-109, 2021. https://doi.org/10.14801/jkiit.2020.18.2.101
  8. I. Goodfellow, Y. Bengio and A. Courville, "Deep Learning," MIT Press, Cambridge, MA, USA, 2016.
  9. K. P. Murphy, "Machine Learning : A Probabilistic Perspective," MIT, MIT Press, Cambridge, MA, USA, 2012.
  10. K. Pawar, R. S. Jalem and V. Tiwari , "Stock Market Price Prediction Using LSTM RNN,"  Emerging Trends in Expert Applications and Security, Advances in Intelligent Systems and Computing, Singapore: Springer, Vol. 841, pp. 493-503, 2019.
  11. X. Song, Y. Liu, L. Xue, J. Wnag, J. Zhang, J. Wang, L. Jiang and Z. Cheng, "Time-Series Well Performance Prediction based on Long Short-Term Memory(LSTM) Neural Network Model," Journal of Petroleum Science and Engineering, Vol. 186, 106682, 2020.
  12. A. Graves, "Generating Sequences With Recurrent Neural Networks," arXiv preprint(https://arxiv.org/pdf/1308.0850), 2014.
  13. H. Yu, Z. Ouyang, W. -P. Zhu, B. Champagne and Y. Ji, "A Deep Neural Network based Kalman Filter for Time Domain Speech Enhancement," 2019 IEEE International Symposium on Circuits and Systems, pp. 1-5, May, 2019.
  14. S. Ouala, R. Fablet, C. Herzet, B. Chapron, A. Pascual, F. Collard and L. Gaultier, "Neural Network based Kalman Filters for the Spatio-Temporal Interpolation of Satellite-Derived Surface Temperature," Remote Sensing, Vol. 10, No. 12, pp. 1-16, 2018. https://doi.org/10.3390/rs10121864
  15. D. Li, J. Zhou and Y. Liu, "Recurrent-Neural-Network-based Unscented Kalman Filter for Estimating and Compensation the Random Drift of MEMS Gyroscopes in Real Time," Mechanical Systems and Signal Processing, Vol. 147, 107057, 2021.
  16. G. Revach, N. Shlezinger, X. Ni, A. Escoriza, L., Van Sloun and Y. C. Eldar, "KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamics," IEEE Transactions on Signal Processing, Vol. 70, pp. 1532-1547, 2022. https://doi.org/10.1109/TSP.2022.3158588
  17. K. Cho, B. Merrienbore, D. Bahdanau, and Y. Bengio, "On the Properties of Neural Machine Translation : Encoder-Decoder Approaches," arXiv preprint(https://arxiv.org/pdf/1409.1259.pdf), 2014.
  18. S. Hochreiter, and J. Schmidhuber, "Long Short-Term Memory," Neural Computation, Vol. 9, No. 8, pp. 1735-1780, 1997. https://doi.org/10.1162/neco.1997.9.8.1735
  19. M. Shirdel, R. Asadi, D. Do and M. Hintlian. "Deep Learning with Kernel Flow Regularization for Time Series Forecasting," arXiv preprint(https://arxiv.org/abs/2109.11649), 2021. 109.11649)
  20. N. Srivastava, G. Hintin, A. Krizhevsky, I. Sutskever and R. Salakhutdinov, "Dropout : A Simple Way to Prevent Neural Networks from Overfitting," The Journal of Machine Learning Research, Vol. 15, No. 1, pp. 1929-1958, 2014.
  21. S. Ynag, X. Yu and Y. Zhou, "LSTM and GRU Neural Network Performance Comparrison Study: Taking Yelp Review Datasets as an Example," 2020 International Workshop on Electronic Communication and Artificial Intelligence(IWECAI), pp. 98-101, 2020.