Implementation of Compressive Sensing

Type: Master's Assignment
Student: Hubert Flisijn
Contacts: Koen Blom
Project: External
Location: Delft, Advanced Developments group, Thales Nederland B.V., The Netherlands

Background

Compressive Sensing (CS) is a new paradigm in sensing (blooming since 2004) that works with a reduced number of measurements for a comparable sensing result. CS offers great potential for reducing complexity and costs, not only in
video cameras and medical sensors but also, in radar systems. There are various applications as well as radar-specific issues that can suit this low-cost paradigm.

Examples of the CS-suitable applications are: array processing, analogue-to-digital conversion, detection, etc. Specific issues are: signal-to-noise ratio, higher resolution, waveforms, data reduction, CS-grid match, implementation, etc.
The well-founded concept of CS has been extensively demonstrated on simulated data. The CS applicability is still to be demonstrated in realistic cases.

Assignment

Thales NL proposes an MSc thesis project whose aim is to investigate
implementation of CS in radar, in a particular case of implementing sparse-signal
estimators on GPU. Simulated as well as real data are to be used to demonstrate
the CS implementation in realistic radar cases of sparse signal reconstruction.
Proposed project planning:

  • studying literature about the state of the art in the CS implementation,
  • developing / implementing the appropriate sparse-signal estimator(s) in CUDA,
  • testing the CS implementation with simulated radar data,
  • evaluating the CS implementation with real radar data, and
  • reporting the CS implementation in radar, in a Master Thesis.
Free Joomla Templates designed by Web Hosting Top