An Improved Stagewise Orthogonal Matching Pursuit-Recursive Algorithm for Power Quality Signal Reconstruction

An improved stagewise orthogonal matching pursuit (iStOMP) algorithm is proposed and used for power quality signal reconstruction in this paper. Firstly, the stagewise orthogonal matching pursuit (StOMP) algorithm is employed to selected atoms to form initial support set. Then the least square method is used to refine the atoms to form a new support set, and updates the residuals accordingly. The newly proposed iStOMP algorithm overcomes the shortcoming of insufficient reconstruction accuracy inherent in the StOMP algorithm and reduces the time consumption. It is because in the second selection the iStOMP algorithm eliminates some suboptimal solutions brought by the first selection and some redundant items in the last iteration. In addition, the iStOMP algorithm retains the characteristics of threshold selection of StOMP algorithm and thus has a strong ability of noise-resilient. Simulation results show that the proposed iStOMP algorithm has higher reconstruction accuracy, higher noise-resilient ability and less time consumption compared with both of the traditional Orthogonal Matching Pursuit-recursive (OMP) and StOMP algorithms. It is thus more suitable for complex power quality signal reconstruction.