Recursive least squares estimation

Abstract

So far, we have considered the least squares solution to a particularly simple es- 3 timation problem in a single unknown parameter. A more general problem is the estimation of the n unknown parameters aj , j = 1, 2, . . . ,n, appearing in a general nth order linear regression relationship of the form, $$ x(k)={a_1}{x_1}(k)+{a_2}{x_2}(k) +\cdots +{a_n}{x_n}(k)$$