基于子空间的正交匹配追踪算法

Research of Orthogonal Matching Pursuit Algorithm on the Base of Subspace

压缩感知理论的提出,给信号处理和信息获取领域带来了划时代的发展。传统重构压缩算法大多都有迭代次数多、运算效率不高且重构效率低等问题。针对该问题,提出了一种根据子空间回溯思想重构出原始信号,并证明了该算法的有效性和重要2个特点:引入回溯思想,重构概率高;计算复杂度低。通过仿真实验与传统的正交跟踪(OMP)算法和子空间(SP)算法进行相关参数比较,验证了该算法在稀疏信号重构研究中具有重要意义。

As the theory of compressed sensing was proposed, it had brought a landmark to the develop- ment of signal processing and obtaining information domains. Most traditional compressed reconstruction al- gorithms had some problems, such as the high iterations, low efficiency of operation and low recovery prob- ability and so on. This paper had proposed backtracking idea to recovery original signal on the base of sub- space, which also proved its availability and two important characteristics： firstly, high recovery probabili- ty because of drawing into backtracking idea, and secondly, the low computational complexity. This paper had compared parameters with traditional Orthogonal Matching Pursuit algorithm and Subspaee Pursuit al- gorithm, and proved its significant importance in the sparse signal recovery domain.