Dynamic Sparse Signal Recovery

T M, Rohan and Detroja, Ketan P and M, Vidyasagar (2018) Dynamic Sparse Signal Recovery. Masters thesis, Indian Institute of Technology Hyderabad.

Text
Thesis_Mtech_EE_4193.pdf - Submitted Version
Restricted to Repository staff only until 2020.
Download (5MB) | Request a copy

Abstract

The thesis was intended to check how the notion of sparsity can be used in control perspective. In the thesis, we present 2 methods for reconstruction of sparse signals with temporal correlation from noisy compressed sensing measurements. This has widespread application in medical imaging, WSNs etc. The proposed methods are based on Kalman Filter formulation. First method ’Re-weighted Sparse Kalman Filter’ which works on the convex representation of Kalman Filter update equation and penalizing it to induce sparsity. In the second method ’Pseudo Observation Approach’ the modified Error covariance and Kalman Filter equation are used along with standard Kalman Filter. Numerical studies where conducted to validate the algorithms.

[error in script]

IITH Creators: