An Improved FOCUSS Algorithm in the Presence of Noise

FOCUSS(Focal Underdetennined System Solver) algorithm is a novel tool for sparse representation and underdetermined inverse problems. To improve the performance of FOCUSS in the presence of noise, we propose an improved FOCUSS algorithm using augmented Lagrange multiplier in this paper. The improved algorithm introduces precise penalty factor with limited value to enhance the precision of signal recovery. For dealing with large scale problem, Conjugate Gradient method is also cooperated to accelerate the convergence speed. Finally, computer simulations illustrate that performance on strengthening the recovery ability of the signal.