شناسايي و كنترل سيگنال هاي سينوسي ميرا با استفاده از شناساننده تطبيقي

چكيده انگليسي :

Identification and Control of Exponentially Damped Sinusoidal Signals Using Adaptive Identifier Mohammad Ali Ghadiri Modarres ma ghadirimodarres@ec iut ac ir Date of Submission 2011 09 11 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran D Supervisor Mohsen Mojiri mohsen mojiri@cc iut ac irAbstractOnline estimation of the parameters of an exponentially damped sinusoidal EDS signal is a classicalproblem in systems theory that has some practical applications Examples include audio signal modeling andsystem identification Audio signals contain relatively fast variations in amplitude and cannot beefficiently modeled as constant amplitude sinusoids In system identification problems thecharacterization of the impulse response of a linear system by a sum of complex exponentials and thenidentifying or approximating the complex amplitudes and the natural frequencies with high degree ofaccuracy is of special importance Parameter estimation of an EDS signal consists of amplitude dampingfactor frequency and phase of the signal Nonlinear dependence of an EDS signal on its frequency anddamping factor makes the estimation of these two parameters a challenging theoretical problem In thisthesis an adaptive algorithm is proposed for the estimation of the frequency and damping factor of an EDSsignal The algorithm is based on an adaptive identifier which is a theoretical tool for the estimation of theparameters of an unknown linear system The governing equations of the proposed algorithm are composedof a second order linear filter and two update laws that provide indirect estimation of the frequency anddamping factor Since the EDS signal is a finite duration signal the parameters which are estimated by thestandard gradient method are not guaranteed to converge to their nominal values To guarantee theconvergence the parameters update laws derived by the gradient algorithm are properly normalized Localasymptotic stability of the proposed algorithm is carried out based on the averaging theory Various aspectsof the proposed algorithm are investigated through extensive computer simulations Specifically a number ofsimulations are conducted to show that the proposed method is appropriate for estimation of the parametersthat are slowly time varying Moreover it is shown that the algorithm is capable of estimating the frequencyand damping factor of an exponentially undamped sinusoidal signal and estimating the frequency of asinusoidal signal The thesis further contributes in presenting an adaptive algorithm for the rejection of theEDS type disturbances affecting a linear and time invariant system For this the proposed algorithm for theestimation of frequency and damping factor of an ESD signal is used The proposed compensator which isbased on the concept of the adaptive feedforward cancelation AFC scheme reproduces the disturbancesignal in the loop The advantage of the proposed compensator is that the frequency and damping factor ofthe disturbance signal is estimated independently of the cancellation scheme for an EDS signal with knownfrequency and damping factor The proposed algorithm is simple in structure and can be easily designed aslong as the plant is known and stable Based on simulation examples the performance of the compensator isevaluated Compared with the internal model principle algorithm the parameters tuning in the proposedmethod is simpler Fast convergence rate of the parameters and rejection of the disturbance signal in shorttime are the other salient features of the system Simulation results confirm the analytical derivations and thedesirable performance of the proposed compensator Key Words Exponentially damped sinusoidal signal Frequency Damping factor Adaptive identifier Averaging theory Adaptive feedforward cancellation

قديري مدرس، محمدعلي

شناسايي و كنترل سيگنال هاي سينوسي ميرا با استفاده از شناساننده تطبيقي

فركانس , ضريب ميرايي , قضيه ي معدل گيري , حذف پيشخوردي تطبيقي