Optimum training design for neural network in synthesis of robust model predictive control

The paper deals with determining the neural network model uncertainty for the purpose of robust controller design. The approach presented in the paper is based on the application of optimum experimental design for the choice of sequences providing the most informative data during the training of neural network. As a criterion quantifying the quality of training process a measure operating on the Fisher Information Matrix related to the estimates of network parameters is used. Then, it is possible to analyze the variance of the predicted network response and estimate how possible variations of parameter values influence the changes observed in the predicted model output. This allows to construct an appropriate cost function for the control system taking into account the model uncertainty and incorporate it into model predictive control scheme.

[1]  B. Bogacka,et al.  Optimum group designs for random-effects nonlinear dynamic processes , 2010 .

[2]  Krzysztof Patan,et al.  Optimal Training Sequences for Locally Recurrent Neural Networks , 2009, ICANN.

[3]  D. Ucinski.  Sensor network scheduling for identification of spatially distributed processes , 2010 .

[4]  Maciej Patan,et al.  Distributed scheduling of sensor networks for identification of spatio-temporal processes , 2012, Int. J. Appl. Math. Comput. Sci..

[5]  Lennart Ljung,et al.  Comparing different approaches to model error modeling in robust identification , 2002, Autom..

[6]  Piotr Tatjewski,et al.  Advanced Control of Industrial Processes: Structures and Algorithms , 2006 .

[7]  Eduardo F. Camacho,et al.  Model Predictive Controllers , 2007 .

[8]  O. Nelles.  Nonlinear System Identification , 2001 .

[9]  Anthony C. Atkinson,et al.  Optimum Experimental Designs , 1992 .

[10]  Krzysztof Patan,et al.  Artificial Neural Networks for the Modelling and Fault Diagnosis of Technical Processes , 2008 .

[11]  David W. Bacon,et al.  Accounting for uncertainty in control-relevant statistics , 2005 .

[12]  XIAOKE YANG,et al.  Fault Tolerant Control Using Gaussian Processes and Model Predictive Control , 2013, 2013 Conference on Control and Fault-Tolerant Systems (SysTol).

[13]  Ian Postlethwaite,et al.  Multivariable Feedback Control: Analysis and Design , 1996 .

[14]  M. Milanese,et al.  Set membership identification of nonlinear systems , 2000, Proceedings of the 39th IEEE Conference on Decision and Control (Cat. No.00CH37187).

[15]  A.G. Alleyne,et al.  A survey of iterative learning control , 2006, IEEE Control Systems.

[16]  Simon Haykin,et al.  Neural networks , 1994 .