Title :
The spectral parameter estimation by using prefiltering and matrix pencil method
Author :
Hu, Fengduo ; Sarkar, T. ; Hua, Yingbo
Author_Institution :
Dept. of Electr. & Comput. Eng., Syracuse Univ., NY, USA
Abstract :
A new approach called the band-pass matrix pencil (BPMP) method for estimating the parameters of exponentially damped or undamped sinusoidal signals is developed. The signal-to-noise ratio (SNR) can be enhanced by using the prefiltering process when prior information about the approximate location and the bandwidth of the signal is available. This information can be obtained by checking the periodogram of the signal. The matrix pencil (MP) method is then applied to the filtered data to estimate frequencies and damping factors of sinusoidal signals. The prefiltering process is not trivial, therefore, the approach presented utilizes the backward process for the IIR filtering and circular convolution for the FIR filtering. Monte Carlo simulations are used to illustrate the performance of the proposed estimation procedure.<>
Keywords :
Monte Carlo methods; filtering and prediction theory; matrix algebra; parameter estimation; spectral analysis; FIR filtering; IIR filtering; Monte Carlo simulations; backward process; band-pass matrix pencil; bandwidth; circular convolution; damping factors; exponentially damped sinusoids; filtered data; frequency estimation; periodogram; prefiltering process; signal location; signal processing; signal-to-noise ratio; spectral analysis; spectral parameter estimation; undamped sinusoidal signals; Bandwidth; Convolution; Damping; Filtering; Finite impulse response filter; Frequency estimation; IIR filters; Parameter estimation; Signal processing; Signal to noise ratio;
Conference_Titel :
Spectrum Estimation and Modeling, 1990., Fifth ASSP Workshop on
Conference_Location :
Rochester, NY, USA
DOI :
10.1109/SPECT.1990.205543