Random partial update sum-squared autocorrelation minimization algorithm for channel shortening (RPUSAM)

Partial updating is an effective method for reducing computational complexity in adaptive filter implementations. In this work, a novel random partial update sum-squared auto-correlation minimization (RPUSAM) algorithm is proposed. This algorithm has low computational complexity whilst achieving improved convergence performance, in terms of achievable bit rate, over a partial update sum-squared auto-correlation minimization (PUSAM) algorithm with a deterministic coefficient update strategy. The performance advantage of the RPUSAM algorithm is shown on eight different carrier serving area test loops (CSA) channels and comparisons are made with the original SAM and the PUSAM algorithms.