"In this case, the right side of the foregoing equation is of the form Hx, where H is a 2-by-S matrix and x is a 2-element column vector as follows: ##EQU1## In accordance with conventional techniques, an estimate zLS of z over the S samples which minimizes the squares of the deviations between the estimate and the true (but a priori unknown) values for the n samples is obtained by performing the f" . . . .