Robustness of the hopfield estimator for identification of dynamical systems

In previous work, a method for estimating the parameters of dynamical systems was proposed, based upon the stability properties of Hopfield networks. The resulting estimation is a dynamical system itself, and the analysis of its properties showed, under mild conditions, the convergence of the estimates towards the actual values of parameters. Also, it was proved that in the presence of noise in the measured signals, the estimation error remains asymptotically bounded. In this work, we aim at advancing in this robustness analysis, by considering deterministic disturbances, which do not fulfill the usual statistical hypothesis such as normality and uncorrelatedness. Simulations show that the estimation error asymptotically vanishes when the disturbances are additive. Thus the form of the perturbation affects critically the dynamical behaviour and magnitude of the estimation, which is a significant finding. The results suggest a promising robustness of the proposed method, in comparison to conventional techniques.

[1]  Gonzalo Joya,et al.  Hopfield Neural Networks for Parametric Identification of Dynamical Systems , 2005 .

[2]  S. Abe.  Theories on the Hopfield neural networks , 1989, International 1989 Joint Conference on Neural Networks.

[3]  Gonzalo Joya,et al.  Dynamical analysis of continuous higher-order hopfield networks for combinatorial optimization. , 2005, Neural computation.

[4]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[5]  J J Hopfield,et al.  Neurons with graded response have collective computational properties like those of two-state neurons. , 1984, Proceedings of the National Academy of Sciences of the United States of America.

[6]  Francisco Sandoval Hernández,et al.  Estimation of the Rate of Detection of Infected Individuals in an Epidemiological Model , 2007, IWANN.

[7]  Teresa Mendonça,et al.  Hopfield neural networks for on-line parameter estimation , 2009, Neural Networks.

[8]  Gonzalo Joya,et al.  Modelling the HIV-AIDS Cuban Epidemics with Hopfield Neural Networks , 2009 .

[9]  Lennart Ljung,et al.  System Identification: Theory for the User , 1987 .

[10]  J. J. Hopfield,et al.  “Neural” computation of decisions in optimization problems , 1985, Biological Cybernetics.

[11]  Mounir Ghogho,et al.  Maximum likelihood estimation of amplitude-modulated time series , 1999, Signal Process..