Nowadays, communication systems require huge amounts of data to be processed. Some examples of these systems include radar systems, video streaming, and many other multimedia applications. These systems require large amounts of bandwidth to satisfy the Nyquist rate. Compressive Sensing is proposed as a way to reduce their bandwidth requirements. Compressive Sensing algorithms are generally implemented at the receiver to reconstruct the original signal from a reduced set of samples. This methodology eliminates data which is relatively insignificant. It possesses the potential to eliminate the use of large bandwidth, cost effective matched filters, and high-frequency analog-todigital converters at the receiver in the case of radar systems. Compressive Sensing is widely used in areas such as Digital Image Processing, Digital Signal Processing, Radars, and Wireless Sensor Networks. This research investigates on three main optimization techniques commonly used in Compressive Sensing: Optimal Matching Pursuit (OMP), Compressive Sampling Matching Pursuit (CSMP) and Stagewise Orthogonal Matching Pursuit (StOMP). These algorithms were implemented and tested on an ARM processor, and on a Field Programmable Gate Array (FPGA). During the first stage of this research, the optimization techniques were implemented in MATLAB. In the second stage, they were implemented on an ARM processor to accelerate their performance. The algorithms show a considerable acceleration on the ARM processor compared to MATLAB. In the final stage of the research, linear algebra operations were implemented on an FPGA to further accelerate their performance. The results show further improvement when part of the code was implemented on an FPGA.

Date of publication

Spring 6-17-2015

Document Type




Persistent identifier


Available for download on Saturday, June 17, 2017