|
Abstract: A method for construction of single resolution and
multiresolution modulating functions is proposed. Based on numerical
integration rule, the structure of modulating filter is obtained. Using
numerical modulating integration filters, an identification model that
is parameterized directly in continuous-time model parameters is given.
By applying the results from discrete-time model identification with the
deriv identification model, a least-squares algorithm and an
instrumental variable(IV) method for identification of continuous-time
dynamic system are formulated. The relationship between the accuracy of
identification and the parameter of modulating filter is investigated,
and some tips about gaussian wavelet modulating function are outlined.
Finally, the methods are applied to identification of dynamic model of
HAGC of a cold mill to confirm the theoretical results.
Key words: Parameter estimation Continuous-time system Modulating
functions Gaussian function Hydraulic AGC
CLC No: TH137
国家自然科学基金(59835170)和湖南省教育厅青年基金(02B032)资助项目. Received 20030211, received
in revised form 20030625
|