Projection-Based and Look-Ahead Strategies for Atom Selection

摘要: In this paper, we improve iterative greedy search algorithms in which atoms are selected serially over iterations, i.e., one-by-one iterations. For serial atom selection, devise two new schemes to select an from a set of potential each iteration. The lead algorithms. both the algorithms, iteration, is found using standard matched filter. case first scheme, propose orthogonal projection strategy that selects atoms. Then, for second look-ahead such selection current iteration has effect on future use requires higher computational resource. To achieve tradeoff between performance and complexity, cascade develop third algorithm. Through experimental evaluations, compare pro posed with existing convex relaxation