MUSE-Fi: Contactless MUti-person SEnsing Exploiting Near-field Wi-Fi Channel Variation

Welcome to MUSE-Fi! This repository contains the code and resources to realize MUSE-Fi, the first Wi-Fi multi-person sensing system with physical separability.

Table of Contents

• Introduction
• Getting Started
• BFI and CSI Parsing and Comparison
• Sparse Recovery Algorithm (SRA)
• Gesture Detection
• Activity Recognition
• CSI Extraction on Smartphones
• License

Introduction

MUSE-Fi is the first Wi-Fi multi-person sensing system with physical separability. This repository contains the code and resources for the following five main parts of MUSE-Fi:

1. BFI and CSI Parsing and Comparison
2. Sparse Recovery Algorithm (SRA)
3. Gesture Detection
4. Activity Recognition
5. CSI Extraction on Smartphones

Whether you're a researcher, developer, or enthusiast, MUSE-Fi opens up possibilities for advanced multi-person sensing applications. This repository provides you with the necessary resources to implement MUSE-Fi effectively.

Getting Started

To get started with MUSE-Fi and use the code to realize the system, follow these steps:

1. Clone this repository to your local machine using git clone https://github.com/uio789bnm/MUSE-Fi/.git.
2. Install Python 3.8 on your computer. You can download the latest version of Python from the official website: python.org.
3. Set up a virtual environment (recommended) using tools like virtualenv or conda. This ensures a clean and isolated environment for your MUSE-Fi implementation. Activate the virtual environment.
4. Download the required datasets. The specific datasets used with MUSE-Fi can be obtained from Dataset Download. Extract the contents of the Data.zip file and make sure the extracted Data folder and the five code folders (gesture, activity, bfi-csi-comparison, SRA, and nexmon-csi) are placed in the same directory.
5. Install the necessary packages such as pytorch 2.0.0, tqdm 4.65.0, numpy 1.23.5, pandas 2.0.3

Now you're ready to implement and use MUSE-Fi!

BFI and CSI Parsing and Comparison

The bfi-csi-comparison folder contains the code for parsing BFI and CSI data in MUSE-Fi. This module also provides functionality for comparing the BFI and CSI data. To utilize this module, follow these steps:

1. Navigate to the bfi-csi-comparison folder: cd bfi-csi-comparison.
2. Ensure your virtual environment is activated.
3. Run the run_Gen_BFICSI_CompareData.m script using appropriate software (e.g., MATLAB).

Sparse Recovery Algorithm (SRA)

The SRA folder contains the code and resources for Sparse Recovery Algorithm of MUSE-Fi. To run the code, follow these steps:

1. Navigate to the SRA folder: cd SRA.
2. Ensure your virtual environment is activated.
3. Run the run_sra_train.sh script: bash run_sra_train.sh.

Gesture Detection

The gesture folder contains the code and resources for gesture recognition using MUSE-Fi. To run the gesture detection classifier, follow these steps:

1. Navigate to the gesture folder: cd gesture.
2. Ensure your virtual environment is activated.
3. Run the main.py script using Python 3.8: python main.py.

Activity Recognition

The activity folder contains the code and resources for activity recognition using MUSE-Fi. To run the activity recognition classifier, follow these steps:

1. Navigate to the activity folder: cd activity.
2. Ensure your virtual environment is activated.
3. Run the main.py script using Python 3.8: python main.py.

CSI Extraction on Smartphones

To perform CSI extraction on smartphones and extend MUSE-Fi's multi-person sensing capabilities to mobile devices, please follow the guidance provided in the external repository Nexmon CSI. The Nexmon CSI repository offers valuable instructions and tools for extracting CSI data on smartphones. As part of the implementation, you may refer to the txt files provided in the nexmon-csi repository as an example.

License

This repository is licensed under the MIT License. For more details, please refer to the LICENSE file.