chapter  5
8 Pages

Eigenvalues of Rx — properties of the error surface

WithAlexander D. Poularikas, Zayed M. Ramadan

Let Rx be an M × M correlation matrix of a WSS discrete-time process computed from a data vector x(n). This symmetric and non-negative matrix (see Problem 5.1.1) can be found using MATLAB as follows: we find the data vector x(n) = [ − 0.4326 − 1.6656 0.1253 0.2877 −1.1465 1.1909 1.1892 −0.0376 0.3273 0.1746 ] using the MATLAB command: x = randn(1, 10). From this data vector we find its autocorrelation function r x(m) = [0.7346 0.0269 –0.1360] using the Book MATLAB function: rx = aasamplebiasedau-toc(x, 3). Next, we obtain the correlation matrix