Premium
Non‐Parametric Identification Method of Volterra Kernels for Nonlinear Systems Excited by Multitone Signal
Author(s) -
Han H. T.,
Ma H. G.,
Tan L. N.,
Cao J. F.,
Zhang J. L.
Publication year - 2014
Publication title -
asian journal of control
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.769
H-Index - 53
eISSN - 1934-6093
pISSN - 1561-8625
DOI - 10.1002/asjc.729
Subject(s) - parametric statistics , nonlinear system , frequency domain , interpolation (computer graphics) , identification (biology) , signal (programming language) , computer science , nonlinear system identification , algorithm , time domain , system identification , excitation , control theory (sociology) , mathematics , engineering , telecommunications , physics , artificial intelligence , data modeling , statistics , botany , control (management) , quantum mechanics , frame (networking) , database , electrical engineering , computer vision , biology , programming language
To solve the problem of Volterra frequency‐domain kernels ( VFKs ) of nonlinear systems, which can be difficult to identify, we propose a novel non‐parametric identification method based on multitone excitation. First, we have studied the output properties of VFKs of nonlinear systems excited by the multitone signal, and derived a formula for identifying VFKs . Second, to improve the efficiency of the non‐parametric identification method, we suggest an increase in the number of tones for multitone excitation to simultaneously identify multi‐point VFKs with one excitation. We also propose an algorithm for searching the frequency base of multitone excitation. Finally, we use the interpolation method to separate every order output of VFK and extract its output frequency components, then use the derived formula to calculate the VFKs . The theoretical analysis and simulation results indicate that the non‐parametric method has a high precision and convenience of operation, improving the conventional methods, which have the defects of being unable to precisely identify VFKs and identification results are limited to three‐order VFK .