A Modeling Method of Switched Systems Based on AP Clustering

Abstract:

The behavior of hybrid systems is determined by combining continuous variables with discrete switching logic. The identification of a hybrid systems aims to find an accurate model of the system's dynamics based on its past inputs and outputs. For a class of hybrid systems in the switched form, a model identification method based on affinity propagation clustering is presented in this paper. In this method, the model identification issue is equivalent to the problems of classification of the cluster data of the system and regression of the classification data. The clustering algorithm is successively applied to group the sampled data into clusters, and the sub-models are trained by least squares support vector machines according to corresponding sub-class samples. The last model of switched systems is obtained by combining each sub-model. Simulation results show the effectiveness and feasibility of the proposed method.

• Periodical: Applied Mechanics and Materials (Volumes 556-562)
• Pages: 3567-3570
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.556-562.3567
• Online since: May 2014
• Authors: Chun Li Xie*, Dan Dan Zhao, Pei Chang Wang
• Keywords: Affinity Propagation Clustering, Least Squares Support Vector Machines, Switched Systems

References

[1] F. Lauer and G. Bloch. Switched and piecewise nonlinear hybrid system identification, In Proc. of the 11th Int. Conf. on Hybrid Systems: Computation and Control (HSCC), St-Louis, MO, USA, vol. 4981 of LNCS (2008), pp.330-343.

DOI: 10.1007/978-3-540-78929-1_24

[2] T. H. Pan, Z. K. Xue, S. Y. Li,. An online multi-model identification algorithm based on subtractive clustering, Acta Automatica Sinica, Vol. 35 (2009), pp.220-224.

DOI: 10.3724/sp.j.1004.2009.00220

[3] B. J. Frey, D. Dueck. Clustering by passing message between data points. Science, Vol. 315 (2007), pp.972-976.

DOI: 10.1126/science.1136800

[4] J. A. K. Suykens, J. Vandewalle. Least squares support vector machine classifiers, Neural Network Letters, Vol. 9 (1999), pp.293-300.

[5] L. L. Van, B. Gérard, L. Fabien. Reduced-Size Kernel Models for Nonlinear Hybrid System Identification, IEEE Transactions on Neural Networks, Vol. 22 (2011), pp.2398-2405.

DOI: 10.1109/tnn.2011.2171361