Design Of Orthogonal Matching Pursuit Algorithm Based On FPGA

Compressed Sensing (CS) is a novel signal acquisition technology that applyingthe compression and gathering simultaneously, which breaks through the traditionalmethod that making the signal sampling before compression. Hence, CS shows thepromising application for the reason that it not only greatly improves the samplingrate, but also reduces the time and storage space. CS is characteristic of easy codingbut complex reconstruction, and studying the reconstruction algorithm is with theimportant significance for the need of real-time application, mainly because of its fastcalculation and high speed implementation, which makes it become the hot topic inthe current field. With the development of FPGA technology, the realization ofhigh-performance digital system based on FPGA gains the extensive attention.The FPGA design and implementation of the orthogonal matching pursuitalgorithm has been investigated in compressed sensing reconstruction. The main workof the paper include：(1) OMP reconstruction algorithm is discussed and analyzed based on studying ofcompressive sensing theory. The issue of matrix inversion module in FPGA designingis clearly discussed after analyzing the OMP reconstruction algorithm.(2) The paper not only discusses the computing matrix inversion which is basedon modified Cholesky decomposition method, but also studies the structural design ofmatrix inversion with FPGA. The programming of the hardware description languageis accomplished in this paper. And it uses third-party software-Modelsim to make thesystem simulation. The simulation results verify the feasibility and effectiveness ofthe design.