GRID matching in Monte Carlo Bayesian compressive sensing

Ioannis Kyriakides, Radmila Pribic, Huseyin Sar, Nuray At

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


Sparse signal reconstruction from compressive measurements assumes a grid of possible support points from which to estimate the signal support set. However, reconstruction of high measurement resolution waveforms is very sensitive to small grid offsets and assuming a fixed grid may result to information loss. On the other hand, identifying sparse elements over a very fine grid to minimize information loss is computationally prohibitive. In this work grid matching is performed via a computationally efficient multi-stage Monte Carlo sampling approach. The multistage sampling method identifies sparse signal elements and chooses the appropriate grid using information from compressively acquired measurements and any prior information on the signal structure. The effectiveness of the method in reconstructing high resolution waveforms, after compressive acquisition, is demonstrated via a simulation study.

Original languageEnglish
Title of host publicationProceedings of the 16th International Conference on Information Fusion, FUSION 2013
Number of pages7
Publication statusPublished - 2013
Event16th International Conference of Information Fusion, FUSION 2013 - Istanbul, Turkey
Duration: 9 Jul 201312 Jul 2013


Other16th International Conference of Information Fusion, FUSION 2013


  • Bayesian compressive sensing
  • grid matching
  • Monte Carlo methods
  • sparse reconstruction


Dive into the research topics of 'GRID matching in Monte Carlo Bayesian compressive sensing'. Together they form a unique fingerprint.

Cite this