“…Given the multivariance Wiener kernels {k (0) 0 , k (1) 1,. , k (2) 2,.,. , ...}, we want to estimate the corresponding Volterra kernels, which are independent of the input variance. Equating (3) and (5), it can be noticed that the two largest order kernels of the WN filter are always equal to the corresponding kernels of the Volterra filter for any input variance σ 2 x , i.e., for r = 0, .., N − 2, s = r, .., N − 1, t = 0, .., N − 1 − s: h 3,t,t+r,t+s = k (3) 3,t,t+r,t+s (6) and for r = 0, .., N − 1, t = 0, .., N − 1 − r:…”