Expanding Window Compressed Sensing for Non-Uniform Compressible Signals.docVIP

Sensors 2012, 12, 13034-13057; doi:10.3390/s121013034 OPEN ACCESS sensors ISSN 1424-8220 /journal/sensors Article Expanding Window Compressed Sensing for Non-Uniform Compressible Signals Yu Liu 1,*, Xuqi Zhu 1, Lin Zhang 1 and Sung Ho Cho 2 1 Key Lab of Universal Wireless Communications, Ministry of Education, Beijing University of Posts and Telecommunications, Beijing 100876, China; E-Mails: safiml@ (X.Z.); zhanglin@ (L.Z.) 2 Department of Electronics and Computer Engineering, Hanyang University, Seoul 133791, Korea; E-Mail: dragon@casp.hanyang.ac.kr * Author to whom correspondence should be addressed; E-Mail: liuy@; Tel.: +86-10-6228-3147; Fax: +86-10-6228-4023. Received: 7 June 2012; in revised form: 28 August 2012 / Accepted: 13 September 2012 / Published: 26 September 2012 Abstract: Many practical compressible signals like image signals or the networked data in wireless sensor networks have non-uniform support distribution in their sparse representation domain. Utilizing this prior information, a novel compressed sensing (CS) scheme with unequal protection capability is proposed in this paper by introducing a windowing strategy called expanding window compressed sensing (EW-CS). According to the importance of different parts of the signal, the signal is divided into several nested subsets, i.e., the expanding windows. Each window generates its own measurements using a random sensing matrix. The more significant elements are contained by more windows, so they are captured by more measurements. This design makes the EW-CS scheme have more convenient implementation and better overall recovery quality for non-uniform compressible signals than ordinary CS schemes. These advantages are theoretically analyzed and experimentally confirmed. Moreover, the EW-CS scheme is applied to the compressed acquisition of image signals and networked data where it also has super