基于同一性的健壮CS分类算法

针对利用压缩感知(CS)进行信号分类识别的问题,提出了一种联合欲分类信号和样本信号的健壮CS分类算法。该方法通过引入“同一性”的概念,克服了信号过完备字典传统构造方式的不足,增强了信号稀疏表示与信号类别间的关联性,提升了基于压缩感知的信号分类算法性能。仿真实验证明了所提方法的正确性,并进一步表明:在非最优过完备字典下,该方法较之传统CS分类算法更具有分类准确度。

A Robust CS Classification Algorithm based on Identity

Aiming at the problem of CS(Compressive Sensing) for signal recognition, a robust CS classification algorithm in combination with test sample and training samples is proposed. By introducing the concept of identity into signal processing,some deficiencies of traditional construction mode for signal over-complete dictionary are overcome,and the correlation of between signal sparse representations and signal classes is also enhanced, thus the performance of classification algorithm based on compressive sensing is improved. Simulation results verify the correctness of the proposed algorithm, and further show that this algorithm enjoys higher classification accuracy than traditional CS classification methods under non-optimal over-complete dictionary condition.