A New Model-Free Predictive Control Method Using Input and Output Data



The purpose of this paper is to present a new predictive control utilizing online data and stored data of input/output of the controlled system. The conventional predictive control methods utilize the mathematical model of the control system to predict an optimal future input to control the system. The model is usually obtained by a standard system identification method from the measured input/output data. The proposed method does not require the mathematical model to predict the optimal future control input to achieve the desired output. This control strategy, called just-in-time, was originally proposed by Inoue and Yamamoto in 2004. In this paper, we proposed a simplified version of the original just-in-time predictive control method.




Khanittha Wongseedakaew and Qi Luo




S. Yamamoto, "A New Model-Free Predictive Control Method Using Input and Output Data", Advanced Materials Research, Vol. 1042, pp. 182-187, 2014


October 2014




* - 通讯作者

[1] Carlos E. Garcia, David M. Prett, and Manfred Morari: Model predictivecontrol: Theory and practice?A survey, Automatica, Vol. 25, Issue 3, p.335–348 (1989).

DOI: https://doi.org/10.1016/0005-1098(89)90002-2

[2] Karl J. Åström and Björn Wittenmark: Adaptive Control: Second Edition, Dover Publication (2008).

[3] A. Stenman: Model-free Predictive Control, Proc. of the 38th IEEE Conferenceon Decision and Control, pp.3712-3717 (1999).

[4] A. Stenman, F. Gustafsson and H. Ljung: Just in time modelsfor dynamical systems, Proc. of the 35th IEEE Conference on Decision andControl, pp.1115-1120(1996).

DOI: https://doi.org/10.1109/cdc.1996.574658

[5] A. Stenman, A. V. Nazin, and F. Gustafsson: Asymptotic propertiesof Just-In-Time models, Preprints of the 11th IFAC Symposiumon System Identification, p.1249–1254 (1997).

[6] M. W. Braun, D. E. Rivera, and A. Stenman: A Model-on-Demand, identification methodology for non-linear processsystems, International Journal of Control, Vol. 74, No. 18, p.1708–1717, (2001).

DOI: https://doi.org/10.1080/00207170110089734

[7] G. Bontempi, M. Birattari and H. Bersini: Lazy Learning forlocal modeling and design; International Journal of Control, Vol. 72, No. 7/8, p.634–658 (1999).

[8] David W. Aha, Dennis Kibler, and Marc K. Albert, Instancebasedlearning algorithms, Machine Learning, Volume 6, Issue1, p.37–66 (1991).

DOI: https://doi.org/10.1007/bf00153759

[9] D. Inoue and S. Yamamoto: An Operation Support Systembased on Database-Driven On-Demand Predictive Control, Proc. of 2004 SICE Annual Conference, p.2024–2027(2004).

[10] D. Inoue and S. Yamamoto: A Memory-Based Predictive ControlApproach to a Braking Assist Problem, Proc. ofSICE Annual Conference 2005, p.3186–3189 (2005).

[11] Q. Zheng and H. Kimura: Locally Weighted Regression Based on k Biparticle Neighbors, Proc. of the 42nd Japan Joint Automatic Control Conference, p.143–144 (1999).

[12] Q. Zheng and H. Kimura: Just-In-Time Modeling for Function Prediction and its Applications, Asian. J. Control, Vol. 3, No. 1, p.35–44 (2001).