Two dimensional adaptive algorithm for system nonlinearly measurements of second order

This work utilizes selected features of the discrete Fourier transform (DFT). The signal of the test is a pseudo-random sequence of ternary levels, which is an anti-symmetrical sequence. The three levels are mapped by an odd sinusoidal signal of proper period. The output of the system contains odd or even harmonics or both, depending on the system kernels. This feature of recognizing the response of the system in the frequency domain is utilized in the separation of these responses. The least mean square algorithm (LMS) is used in this work for estimating both the first and second kernels after some modification and the error functions are presented as a convergence surface.