Wi-Fi multipath parameter estimation

This repository contains the reference code for the article ''Wi-Fi Multi-Path Parameter Estimation for Sub-7 GHz Sensing: A Comparative Study''.

If you find the project useful and you use this code, please cite our article:

@inproceedings{meneghello2023wifimultipath,
  author={Meneghello, Francesca and Blanco, Alejandro and Cusano, Antonio and Widmer, Joerg and Rossi, Michele},
  booktitle={Proc. of IEEE WiMob}, 
  title={{Wi-Fi Multi-Path Parameter Estimation for Sub-7 GHz Sensing: A Comparative Study}}, 
  year={2023},
  address={Montreal, Canada}
  }

How to use

If you want to replicate the quantitative evaluation in the paper above you need to generate simulated channel frequency response data using the data_generation.m code as explained in the first step.

If you want to use the multi-path parameter estimators on your own data skip the generation of the simulated files and directly use the estimators.

Simulated data generation

To generate the channel data for the evaluation of the multi-path parameter estimation algorithms use:

matlab Matlab_code/data_generation.m

The data is generated considering a 1 X 4 MIMO system, i.e., one transmitter antenna and four receiver antennas, operating on a 80 MHz channel with 256 OFDM sub-channels and with 15 dB of SNR.

Multi-path parameter estimation approaches

The following pipelines take as input a file containing channel frequency response (CFR) data and provide as output the estimated parameters of the multi-path components.

mD-Track pipeline

python Python_code/method_mDtrack.py <'number of spatial streams'> <'number of cores'> <'name of the directory of data'> <'starting of the name of the file'> <'delta ToA for grid search in multiples of 10^-11'>

e.g., python Python_code/method_mDtrack.py 1 4 ../simulation_files/change_delay_aoa/ simulation_artificial_grid_delaydiff_1e-09_aoadiff_4 --delta_t 25

The script uses the following utility functions: build_aoa_matrix, build_toa_matrix in utilityfunct_aoa_toa_doppler.py and md_track_2d in utilityfunct_md_track.py.

Compressed sensing pipeline (IHT w/o refinement, IHT with refinement, OMP, LASSO)

python Python_code/method_compressed_sensing.py <'start index for processing'> <'end index for processing'> <'step length'> <'optimization method (among iht_noref, iht, omp, lasso)'> <'name of the directory'> <'starting of the name of the file'>

e.g., python Python_code/method_compressed_sensing_simulation.py 0 -1 1 iht ../simulation_files/change_delay_aoa/ simulation_artificial_grid_delaydiff_1e-09_aoadiff_4

The script uses the following utility functions: build_toa_matrix, build_toa_aoa_matrix in utilityfunct_aoa_toa_doppler.py, and the functions in utilityfunct_optimization_routines.py and utilityfunct_optimization.py. Note that the LASSO routine requires the OSQP solver.

SpotFi pipeline

matlab Matlab_code/method_SpotFi.m

The script uses the functions inside the functions folder.

UbiLocate pipeline

matlab Matlab_code/method_UbiLocate_2D.m

The script uses the functions inside the functions folder.

Performance evaluation

The following pipelines analyze the results obtained through the multi-path parameter estimators and compute the average performance.

For mD-Track, IHT w/o refinement, IHT with refinement, OMP, LASSO

python Python_code/analysis_from_python_methods.py <'step for the AoA of first path'> <'step for the AoA of second path'> <'maximum number of paths detected'> <'optimization method (among iht_noref, iht, omp, lasso)'> <'name of the directory'> <'starting of the name of the file'>

e.g., python Python_code/analysis_from_python_methods.py 1 2 2 mdTrack ../simulation_files/change_delay_aoa/ simulation_artificial_delayobst_1.86e-08

For UbiLocate and SpotFi

python Python_code/analysis_from_matlab_methods.py <'step for the AoA of first path'> <'step for the AoA of second path'> <'maximum number of paths detected'> <'optimization method (among iht_noref, iht, omp, lasso)'> <'name of the directory'> <'starting of the name of the file'> <'whether the amplitudes are in dB (default 0, i.e., not dB)'>

e.g., python Python_code/analysis_from_matlab_methods.py 1 2 2 ubilocate ../simulation_files/change_delay_aoa/ simulation_artificial_delayobst_1.86e-08 --dB 0

For spotfi use --dB 1.

Plotting the results

python Python_code/plot_combined.py <'start for ToA'> <'end for Toa'> <'step for ToA'> <'step for AoA'> <'number of AoA (default 180)'>

e.g., python Python_code/plot_combined.py 1.02e-08 2e-8 2e-10 1

Python and relevant libraries version

Python >= 3.10.9 Numpy >= 1.23.5
Scipy = 1.9.3
Scikit-learn = 1.2.0
OSQP >= 0.6.1

Contact

Francesca Meneghello meneghello@dei.unipd.it github.com/francescamen