A blind lag-hopping adaptive channel shortening algorithm based upon squared auto-correlation minimization (LHSAM)

Recent analytical results due to Walsh, Martin and Johnson showed that optimizing the single lag autocorrelation minimization (SLAM) cost does not guarantee convergence to high signal to interference ratio (SIR), an important metric in channel shortening applications. We submit that we can overcome this potential limitation of the SLAM algorithm and retain its computational complexity advantage by minimizing the square of single autocorrelation value with randomly selected lag. Our proposed lag-hopping adaptive channel shortening algorithm based upon squared autocorrelation minimization (LHSAM) has, therefore, low complexity as in the SLAM algorithm and, more importantly, a low average LHSAM cost can guarantee to give a high SIR as for the SAM algorithm. Simulation studies are included to confirm the performance of the LHSAM algorithm.