WiFi_6GHz_Dataset / README.md
WIFI6GHZDATASET's picture
Update README.md
0bacfcc verified
# Wi6CSI Dataset Overview
**Wi6CSI** is a comprehensive Wi-Fi sensing dataset collected in the **6GHz** band. It offers a rich, high-resolution resource for researchers
## Key Features
- **Scenarios Covered:**
- **Single User:** Data captured with one user performing various activities.
- **Multi User:** Data where multiple users are present simultaneously.
- **People Counting:** Scenarios designed to count individuals within a specific area.
- **Collection Locations:**
Data has been collected across **three distinct locations**, ensuring diverse environmental conditions and variations in the wireless channel.
- **Data Volume and Resolution:**
- The dataset holds approximately **1.5TB** of data.
- Each Wi-Fi packet includes **8100 subcarriers**, offering a level of detail that supports high-resolution sensing and analysis.
- **Wi-Fi Band:**
- Utilizes the **6GHz** frequency band, providing enhanced granularity and improved sensing accuracy compared to lower frequency bands.
# Wi6CSI Sample Dataset
This repository contains a sample subset of the complete Wi6CSI dataset containing just a few single user activities. The sample is provided to demonstrate the dataset structure and typical processing steps for research in Wi-Fi Channel State Information (CSI) analysis.
If you are interested in obtaining access to the complete Wi6CSI dataset or have any questions, please contact: [e0426004@u.nus.edu](mailto:e0426004@u.nus.edu)
## Dataset Contents
- **.mp4 files:**
These files are Pose3D videos that capture 3D representations of subjects during data collection. They can be used for visual analysis, activity recognition, or correlating motion with CSI data.
- **.mat files:**
These MATLAB files contain the Wi-Fi CSI values. The numerical data stored in these files represent the channel state information of the wireless signal, useful for signal processing or machine learning applications.
- **.csi files:**
These files hold raw Wi-Fi CSI data. They are not immediately usable and must be parsed using the [picoscenes toolbox](https://ps.zpj.io/matlab.html) to extract the relevant features for analysis.
## How to View and Process Each File Type
### 1. `.mp4` Files
- **Description:**
These files contain Pose3D videos.
- **How to View:**
Open them in any standard media player (e.g., VLC, Windows Media Player, QuickTime) by double-clicking the file or dragging it into the player window.
### 2. `.csi` Files
- **Description:**
These files store raw Wi-Fi CSI data.
- **How to Open/Parse:**
- **Step 1:** Install the PicoScenes MATLAB Toolbox.
- **Step 2:** Open MATLAB and navigate to the directory containing the `.csi` file.
- **Step 3:** Load or parse the file using the toolbox’s function. Found under the PicoScenes-MATLAB-Toolbox-Core folder
- **Step 4:** The raw CSI values can then be accessed within the returned data structure for further analysis.
### 3. `.mat` Files
- **Description:**
These files contain Wi-Fi CSI values in MATLAB’s native format.
- **How to View:**
- **Option 1:** In MATLAB, double-click the `.mat` file in the Current Folder panel.
- **Option 2:** Programmatically load the file by running. For example:
```matlab
load('A12_P5_L3_T1.mat');
```
After loading, variables such as `data` (often containing a field like `data.CSI.CSI`) will appear in your MATLAB workspace for exploration.
The data.CSI.CSI will then contain all raw complex CSI values.
## Single User Activity Recognition Structure
| Activity |
|---------------------------------|
| Empty Room (A1) |
| Standing Still (A2) |
| Sitting Still (A3) |
| Lying Down Still (A4) |
| Clap (A5) |
| Wave (A6) |
| Arm Circles (A7) |
| Bicep Curls (A10) |
| Toe Touches (A8) |
| Squats (A9) |
| Walk (Perpendicular) (A11) |
| Walk (Parallel) (A12) |
| Jump (A13) |
| Location |
|---------------------------------|
| Room (L1) |
| Extended Room (L2) |
| Meeting Room (L3) |
| Trial Number |
|---------------------------------|
| Trial 1 (T1) |
| Trial 2 (T2) |
| Trial 3 (T3) |
| Trial 4 (T4) |
| Trial 5 (T5) |
| Participant Number |
|---------------------------------|
| Participant 1 (P1) |
| Participant 2 (P2) |
| Participant 3 (P3) |
| Participant 4 (P4) |
| Participant 5 (P5) |