Highly efficient surface-search algorithm for adaptive FIR filtering

This paper proposes a highly efficient Surface search algorithm (SSA) for adaptive FIR filtering. The proposed algorithm depends on the highly efficient calculation of the Fast gain vectors (FGA) defined in Ref.[14], and the rank-one updating formula of the correlation matrix of the input vector sequence. The computational complexity of this algorithm is 6M + 10MAD's (which stands for the number of multiplies and divides), and lower than that of the fastest recursive least squares transversal filters for adaptive FIR filtering. The global convergence of the proposed algorithm is studied by Lyapunov indirect method. The performances of the relative algorithms are shown via computer simulations. The long-term numerical stability of the SSA is also verified via simulations.