Online Compressive Covariance Sensing and Its Applications

Abstract

Several sampling strategies, which compress an analog signal and recover it, are attracting attention. Representatively, compressive sensing (CS) can restore a sparse signal from its compressed one sampled at a lower rate than the Nyquist rate. However, CS is not appropriate to non-sparse signals. Recently developed compressive covariance sensing (CCS) methods have received greㅁd simulation results showed that the proposed methods have high covariance tracking performance as well as accurate restoration. I applied the online CCS technique to real-world applications (healthcare and human computer interface) and compared it with previous methods such as original CCS, CS, and wavelet compression.