SFWCMF

Code for source localization with the covariance matrix fitting methods and Sliding Frank-Wolfe

Code and data to reproduce the results of Gridless covariance matrix fitting methods for three dimensional acoustical source localization

This code is not maintained. See acosolo for python and matlab implementations.

Figures

The proposed methods (CMF and COMET with Sliding Frank-Wolfe) are compared with MUSIC, CLEAN-SC, HR-CLEAN-SC, OMP, and differential evolution (see references in the paper)

• FIG_exp.m: experimental results (noiseless and noisy versions)
• FIG_SNR.m: performance in function of the signal to noise ratio
• FIG_K.m: performance in function of the frequency
• FIG_SNAPS.m: performance in function of the number of snapshots
• FIG_RESOL.m, FIG_RESOL_var.m: comparison of the resolution

Scripts

Most of the scripts are for internal use or for comparison purposes.

The main scripts are:

• sfw_cmf.m: SFW algorithm for the CMF problem
• sfw_comet1.m and sfw_comet2.m : SFW algorithm for the two COMET variants