Adaptive Regularized FOCUSS Algorithm

The sparse solutions obtained via the regularized version of FOCUSS (focal underdetermined system solver) are governed by the choice of a proper regularization parameter. Based on the principle of adaptive regularization algorithm (ARA) and regularized FOCUSS (R-FOCUSS) algorithm, this paper proposes an adaptive regularized FOCUSS (AR-FOCUSS) algorithm for solving an underdetermined linear system. Simulation results show that AR-FOCUSS is less sensitive to the choice of regularization parameter than R-FOCUSS, and the former outperforms the latter under the same conditions.