Reduced order RLS polynomial predistortion

In this paper, a modified recursive least square (RLS) based polynomial predistortion algorithm is proposed by reducing the order of the polynomials used in the RLS algorithm. The basic idea of the reduced order RLS based polynomial predistortion algorithm is to approximate the target function with two simpler functions (with a smaller number of coefficients) instead of a single function. A complex function is decomposed into two simpler functions so that we can compute the RLS algorithm with much less complexity. The proposed reduced order RLS polynomial predistortion shows superior performance compared to the conventional RLS polynomial predistortion with the same number of coefficients or with the same computational complexity.