Permutation enhanced parallel reconstruction for compressive sampling

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Researchers

Research units

  • University of Washington
  • University of Alberta

Abstract

In this paper, a simple but efficient permutation enhanced parallel reconstruction architecture for compressive sampling (CS) is proposed. In this architecture, a measurement matrix is constructed from a block-diagonal sensing matrix, the sparsifying basis of the target signal, and a pre-defined permutation matrix. In this way, the projection of the signal onto the sparsifying basis can be divided into several segments and all segments can be reconstructed in parallel. Thus, the computational complexity and the time for reconstruction can be reduced significantly. With a good permutation matrix, the error performance of the proposed method can be improved compared with the option without permutation. The proposed method can be used in applications where the computational complexity and time for reconstruction are crucial evaluation criteria and centralized sampling is acceptable. Simulation results show that the proposed method can achieve comparable results to the centralized reconstruction methods (i.e., standard CS and distributed CS), while requiring much less reconstruction time.

Details

Original languageEnglish
Title of host publication2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2015
Publication statusPublished - 14 Jan 2016
MoE publication typeA4 Article in a conference publication
EventIEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing - Cancun, Mexico
Duration: 13 Dec 201516 Dec 2015
Conference number: 6
http://inspire.rutgers.edu/camsap2015/

Workshop

WorkshopIEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing
Abbreviated titleCAMSAP
CountryMexico
CityCancun
Period13/12/201516/12/2015
Internet address

    Research areas

  • Permutation, parallel algorithms, signal reconstruction, compressive sampling

ID: 1951227