| Popis: |
The accuracy and resolution of remote sensing instruments employing radio-wave techniques can be increased by applying methods of estimation of the model parameters. On the basis of a priori information on the structure of signals, effective computing algorithms for the construction of models of the investigated process are used. The data model is assumed to consist of a limited number of plane waves (oscillations) embedded in noise, and the signal-to-noise ratio q>1. Methods of spectral estimation based on Levinson's, Berg's, Cholesky's and least squares algorithms with the solution of the system of linear equations by Gauss' method are used in the simulations. Applying different methods of autoregressive parameter estimation, computer simulations are conducted at various values of q and for a number of waves (oscillations). The effects of the order of autoregressive models on the results of the evaluation are investigated, as well as the effects of the length of the time series. |
