A single-precision compressive sensing signal reconstruction engine on FPGAs

Fengbo Ren, Richard Dorrace, Wenyao Xu, Dejan Markovic

Research output: Chapter in Book/Report/Conference proceedingConference contribution

19 Citations (Scopus)

Abstract

Compressive sensing (CS) is a promising technology for the low-power and cost-effective data acquisition in wireless healthcare systems. However, its efficient realtime signal reconstruction is still challenging, and there is a clear demand for hardware acceleration. In this paper, we present the first single-precision floating-point CS reconstruction engine implemented a Kintex-7 FPGA using the orthogonal matching pursuit (OMP) algorithm. In order to achieve high performance with maximum hardware utilization, we propose a highly parallel architecture that shares the computing resources among different tasks of OMP by using configurable processing elements (PEs). By fully utilizing the FPGA recourses, our implementation has 128 PEs in parallel and operates at 53.7 MHz. In addition, it can support 2x larger problem size and 10x more sparse coefficients than prior work, which enables higher reconstruction accuracy by adding finer details to the recovered signal. Hardware results from the ECG reconstruction tests show the same level of accuracy as the double-precision C program. Compared to the execution time of a 2.27 GHz CPU, the FPGA reconstruction achieves an average speed-up of 41x.

Original languageEnglish (US)
Title of host publication2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings
PublisherIEEE Computer Society
DOIs
StatePublished - 2013
Externally publishedYes
Event2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Porto, Portugal
Duration: Sep 2 2013Sep 4 2013

Other

Other2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013
CountryPortugal
CityPorto
Period9/2/139/4/13

Fingerprint

Signal Reconstruction
Signal reconstruction
Compressive Sensing
Field Programmable Gate Array
Field programmable gate arrays (FPGA)
Engine
Engines
Matching Pursuit
Hardware
Parallel architectures
Hardware Acceleration
Processing
Electrocardiography
Computer hardware
Program processors
Parallel Architectures
Data acquisition
Floating point
Data Acquisition
Healthcare

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

Ren, F., Dorrace, R., Xu, W., & Markovic, D. (2013). A single-precision compressive sensing signal reconstruction engine on FPGAs. In 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings [6645574] IEEE Computer Society. https://doi.org/10.1109/FPL.2013.6645574

A single-precision compressive sensing signal reconstruction engine on FPGAs. / Ren, Fengbo; Dorrace, Richard; Xu, Wenyao; Markovic, Dejan.

2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings. IEEE Computer Society, 2013. 6645574.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ren, F, Dorrace, R, Xu, W & Markovic, D 2013, A single-precision compressive sensing signal reconstruction engine on FPGAs. in 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings., 6645574, IEEE Computer Society, 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013, Porto, Portugal, 9/2/13. https://doi.org/10.1109/FPL.2013.6645574
Ren F, Dorrace R, Xu W, Markovic D. A single-precision compressive sensing signal reconstruction engine on FPGAs. In 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings. IEEE Computer Society. 2013. 6645574 https://doi.org/10.1109/FPL.2013.6645574
Ren, Fengbo ; Dorrace, Richard ; Xu, Wenyao ; Markovic, Dejan. / A single-precision compressive sensing signal reconstruction engine on FPGAs. 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings. IEEE Computer Society, 2013.
@inproceedings{0828fd39df78439c81e4fd560f4390a3,
title = "A single-precision compressive sensing signal reconstruction engine on FPGAs",
abstract = "Compressive sensing (CS) is a promising technology for the low-power and cost-effective data acquisition in wireless healthcare systems. However, its efficient realtime signal reconstruction is still challenging, and there is a clear demand for hardware acceleration. In this paper, we present the first single-precision floating-point CS reconstruction engine implemented a Kintex-7 FPGA using the orthogonal matching pursuit (OMP) algorithm. In order to achieve high performance with maximum hardware utilization, we propose a highly parallel architecture that shares the computing resources among different tasks of OMP by using configurable processing elements (PEs). By fully utilizing the FPGA recourses, our implementation has 128 PEs in parallel and operates at 53.7 MHz. In addition, it can support 2x larger problem size and 10x more sparse coefficients than prior work, which enables higher reconstruction accuracy by adding finer details to the recovered signal. Hardware results from the ECG reconstruction tests show the same level of accuracy as the double-precision C program. Compared to the execution time of a 2.27 GHz CPU, the FPGA reconstruction achieves an average speed-up of 41x.",
author = "Fengbo Ren and Richard Dorrace and Wenyao Xu and Dejan Markovic",
year = "2013",
doi = "10.1109/FPL.2013.6645574",
language = "English (US)",
booktitle = "2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - A single-precision compressive sensing signal reconstruction engine on FPGAs

AU - Ren, Fengbo

AU - Dorrace, Richard

AU - Xu, Wenyao

AU - Markovic, Dejan

PY - 2013

Y1 - 2013

N2 - Compressive sensing (CS) is a promising technology for the low-power and cost-effective data acquisition in wireless healthcare systems. However, its efficient realtime signal reconstruction is still challenging, and there is a clear demand for hardware acceleration. In this paper, we present the first single-precision floating-point CS reconstruction engine implemented a Kintex-7 FPGA using the orthogonal matching pursuit (OMP) algorithm. In order to achieve high performance with maximum hardware utilization, we propose a highly parallel architecture that shares the computing resources among different tasks of OMP by using configurable processing elements (PEs). By fully utilizing the FPGA recourses, our implementation has 128 PEs in parallel and operates at 53.7 MHz. In addition, it can support 2x larger problem size and 10x more sparse coefficients than prior work, which enables higher reconstruction accuracy by adding finer details to the recovered signal. Hardware results from the ECG reconstruction tests show the same level of accuracy as the double-precision C program. Compared to the execution time of a 2.27 GHz CPU, the FPGA reconstruction achieves an average speed-up of 41x.

AB - Compressive sensing (CS) is a promising technology for the low-power and cost-effective data acquisition in wireless healthcare systems. However, its efficient realtime signal reconstruction is still challenging, and there is a clear demand for hardware acceleration. In this paper, we present the first single-precision floating-point CS reconstruction engine implemented a Kintex-7 FPGA using the orthogonal matching pursuit (OMP) algorithm. In order to achieve high performance with maximum hardware utilization, we propose a highly parallel architecture that shares the computing resources among different tasks of OMP by using configurable processing elements (PEs). By fully utilizing the FPGA recourses, our implementation has 128 PEs in parallel and operates at 53.7 MHz. In addition, it can support 2x larger problem size and 10x more sparse coefficients than prior work, which enables higher reconstruction accuracy by adding finer details to the recovered signal. Hardware results from the ECG reconstruction tests show the same level of accuracy as the double-precision C program. Compared to the execution time of a 2.27 GHz CPU, the FPGA reconstruction achieves an average speed-up of 41x.

UR - http://www.scopus.com/inward/record.url?scp=84898645166&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898645166&partnerID=8YFLogxK

U2 - 10.1109/FPL.2013.6645574

DO - 10.1109/FPL.2013.6645574

M3 - Conference contribution

BT - 2013 23rd International Conference on Field Programmable Logic and Applications, FPL 2013 - Proceedings

PB - IEEE Computer Society

ER -