Identification of System with Non-stationary Signal Using Modified Wilcoxon Approach

Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 199)


Non-stationary random signals exhibit time-dependent characteristics and require proper models and corresponding identification methods. The focus is on identification method. We study system identification of the non-stationary parameters in this task. In this paper, the problem of non-causal identification of non-stationary, linear stochastic systems has been considered. A robust system identification approach adapted to chirp signals is proposed. An asymptotically unbiased estimate for the system’s transfer function is analyzed. We show that compared to a competing non-stationarity based method, a significantly smaller error variance is achieved and generally shorter observation intervals are required. The adaptive method used here, is Wilcoxon approach based. Also the comparison have been done with Sign WLMS and Sign sign WLMS methods as the modified technique. In case of a time-varying system, faster convergence and higher reliability of the system identification are obtained. The results confirm the advantages of proposed approach. The resulting parallel estimation scheme automatically adjusts its smoothing parameters to the unknown, and possibly time-varying, rate of non-stationarity of the identified system.


System Identification non-stationary signal adaptive signal processing chirp impulse response Wilcoxon norm Sign Wilcoxon Technique 


  1. 1.
    Iem, B.G., Papandreou-Suppappola, A., Faye Boudreaux-Bartels, G.: Wideband Weyl symbols for dispersive timevarying processing of systems and random signals. IEEE Transactions on Signal Processing 50(5), 1077–1090 (2002)CrossRefMathSciNetGoogle Scholar
  2. 2.
    Jiang, Y., Papandreou-Suppappola, A.: Discrete timefrequency characterizations of dispersive linear time-varying systems. IEEE Transactions on Signal Processing 55(5), 2066–2076 (2007)CrossRefMathSciNetGoogle Scholar
  3. 3.
    Papandreou-Suppappola, A., Ioana, C., Zhang, J.J.: Timescale and dispersive processing for wideband time-varying channels. In: Hlawatsch, F., Matz, G. (eds.) Wireless Communications over Rapidly Time-Varying Channels. Academic Press (2011)Google Scholar
  4. 4.
    Gerencsér, L., Hjalmarsson, H., Mårtensson, J.: Identification of ARX systems with non-stationary inputs and asymptotic analysis with application to adaptive input design. Automatica 45, 623–633 (2009)CrossRefMATHGoogle Scholar
  5. 5.
    Niedźwiecki, M.: Identification of time-varying processes. Wiley, New York (2000)Google Scholar
  6. 6.
    Niedźwiecki, M., Gackowski, S.: On noncausal weighted least squares identification of nonstationary stochastic systems. Automatica 47, 2239–2244 (2011)CrossRefMATHGoogle Scholar
  7. 7.
    Chen, B., Hu, J., Li, H., Zengqi: A Joint Stochastic Gradient Algorithm and its application to system identification with RBF Networks. In: Sun Proceedings of the 6th World Congress on Intelligent Control and Automation, China, June 21-23 (2006)Google Scholar
  8. 8.
    Dash, S., Mohanty, M.N.: A Comparative Analysis for WLMS Algorithm in System Identification. In: IEEE Conf. ICECT, Kanyakumari, April 6-7 (2012)Google Scholar
  9. 9.
    Dash, S., Mohanty, M.N.: Analysis of Outliers in System identification using WLMS Algorithm. In: IEEE International Conference on Computing, Electronics and Electrical Technologies, ICCEET, Kanyakumari, March 21-22, pp. 802–806 (2012)Google Scholar
  10. 10.
    Shaik, R.A., Reddy, D.V.R.K.: Noise cancellation in ECG signals using normalized Sign-Sign LMS algorithm. In: IEEE International Symposium on Signal Processing and Information Technology, ISSPIT, December 14-17 (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  1. 1.ITERSiksha ‘O’ Anusandhan UniversityBhubaneswarIndia

Personalised recommendations