CHAPTER 30 Recursive Least-Squares
Now that we have studied in some detail the least-squares problem, we proceed to derive an algorithm for updating its solution. The resulting recursion will be referred to as the Recursive Least-Squares (RLS) algorithm and it will form the basis for most of our discussions in the future chapters.
30.1 MOTIVATION
Given an N × 1 measurement vector y, an N × M data matrix H and an M × M positive-definite matrix Π, we saw in Sec. 29.7 that the M × l solution to the following regularized least-squares problem:
is given by
where, in comparison with (29.28), we are assuming
for simplicity of presentation. The arguments would apply equally well to the case
— see the remark after Lemma 30.1.
We denote the individual entries of y by {d(i)}, and the individual rows of H by {ui}, say,

so that the solution
in (30.2) is determined by data {d(
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access