HTTPS SSH

1. Introduction

This folder contains the code to reproduce the numerical results reported in:

``Super-resolution radar'', by R. Heckel, V.I. Morgenshtern, M. Soltanolkotabi

Specifically, the folder contains code to reproduce Figures 3, 4, 5, and 7, by running the corresponding instructions. For example, make fig4 will run the code to generate the data for Figure 4, stores the data in the ./dat folder, and generates the file ./tex/figure4.pdf containing the figure.

2. Requirements

Running the code requires installations of Python, Matlab, Make, and Pdflatex.

The code for Figures 3 and 7 is written in Python. To solve the corresponding l_1-regularized least-squares problem, we use the SPGL1 solver, described in:

E. van den Berg and M. P. Friedlander, "Probing the Pareto frontier for basis pursuit solutions", SIAM J. on Scientific Computing, 31(2):890-912, 2008.

The SPGL1 solver can be obtained from: git clone https://github.com/drrelyea/SPGL1_python_port.git

The code for Figures 4 and 5 is written in Matlab. To solve the corresponding SDP, we use CVX, which can be obained from: http://cvxr.com/cvx/

3. Quick start

  • Obtain the SPGL1 solver by running the following command in a shell: git clone https://github.com/drrelyea/SPGL1_python_port.git
  • Install CVX in Matlab
  • Running: make all generates all figures, in low resolution.
  • Run:
    • make fig3 # to generate a low resolution version of Figure 3
    • make fig3-orig # to generate the original Figure 3
    • make fig4 # to generate Figure 4 (with N=5 instead N=8)
    • make fig5 # to generate Figure 5
    • make fig7 # to generate a low resolution version of Figure 7
    • make fig7-orig # to generate the original Figure 7

4. Licence

All files are provided under the terms of the Apache License, Version 2.0, see the included file "apache_licence_20" for details.