Compressive sensing reconstruction algorithms using partially correct signal information

Compressive Sensing (CS) ensures the reconstruction of a sparse signal from a set of linear measurements that are fewer compared to the signal length. The sparse signal can be reconstructed using a convex relaxation algorithm such as Basis Pursuit (BP) or a Greedy Pursuit (GP) such as Backtrackin...

Full description

Saved in:
Bibliographic Details
Main Author: Sekar Sathiya Narayanan
Other Authors: Anamitra Makur
Format: Theses and Dissertations
Language:English
Published: 2017
Subjects:
Online Access:http://hdl.handle.net/10356/69572
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Compressive Sensing (CS) ensures the reconstruction of a sparse signal from a set of linear measurements that are fewer compared to the signal length. The sparse signal can be reconstructed using a convex relaxation algorithm such as Basis Pursuit (BP) or a Greedy Pursuit (GP) such as Backtracking Matching Pursuit (BMP). If some information regarding the signal support (non-zero locations) is available in the form of Partially Known Support (PKS), the same sparse signal can be recovered with higher accuracy. However, the size and accuracy of the PKS varies depending upon the signal model and characteristics. A generic PKS based reconstruction algorithm might work well in a particular scenario but fail in another. This thesis focuses on developing PKS based reconstruction algorithms for different scenarios wherein they make effective use of the available PKS.