LoFi: Vision-Aided Label Generator for Wi-Fi Localization and Tracking

Paper: LoFi: Vision-Aided Label Generator for Wi-Fi Localization and Tracking Code: https://github.com/RS2002/LoFi

Abstract

Data-driven Wi-Fi localization and tracking have shown great promise due to their lower reliance on specialized hardware compared to model-based methods. However, most existing data collection techniques provide only coarse-grained ground truth or a limited number of labeled points, significantly hindering the advancement of data-driven approaches. While systems like lidar can deliver precise ground truth, their high costs make them inaccessible to many users. To address these challenges, we propose LoFi, a vision-aided label generator for Wi-Fi localization and tracking. LoFi can generate ground truth position coordinates solely from 2D images, offering high precision, low cost, and ease of use. Utilizing our method, we have compiled a Wi-Fi tracking and localization dataset using the ESP32-S3 and a webcam. The code and dataset of this paper are available at this https URL .

Dataset Structure

Data Instances

The dataset includes raw CSI data in .csv files and processed data in .pkl files. Each raw CSI instance is a .csv file with the following columns:

• seq: Row number of the entry.
• timestamp: UTC+8 time of data collection.
• local_timestamp: ESP32 local time.
• rssi: Received Signal Strength Indicator.
• data: CSI data with 104 numbers representing 52 subcarriers, where each subcarrier's complex CSI value is computed as a[2i] + a[2i+1]j.
• Other columns: Additional ESP32 device information (e.g., MAC, MCS details).

The processed data (./data/processed_data/wiloc.pkl) includes:

• CSI data, position coordinates, person IDs, and timestamps, with -1000 indicating packet loss positions.

Data Fields

Field Name	Description
seq	Row number of the entry (raw CSI data)
timestamp	UTC+8 time of data collection (raw CSI data)
local_timestamp	ESP32 local time (raw CSI data)
rssi	Received Signal Strength Indicator (raw CSI data)
data	CSI data (104 numbers, representing 52 subcarriers as complex values) (raw CSI data)
Other columns	Additional ESP32 metadata (e.g., MAC address, MCS details) (raw CSI data)
Position coordinates	2D location coordinates (processed data, wiloc.pkl)
Person ID	Identifier for the person (processed data, wiloc.pkl)

Data Splits

The dataset is organized as follows:

• Raw CSI Data (./data/csi_data.zip): Organized by person name, with each .csv file containing CSI data for a specific person.
• Processed Data (./data/processed_data.zip):
  • csi_data.pkl: Extracted original CSI data.
  • data_sequence.pkl: CSI data split by person.
  • gt_data.pkl: Position coordinates extracted from images.
  • wiloc.pkl: Combined CSI data, ground truth coordinates, and person IDs.

Dataset Creation

Curation Rationale

The dataset was created to support research on WiFi-based human location, tracking, and identification using low-cost ESP32-S3 devices, enhanced by vision-aided label generation, enabling applications in indoor navigation and smart environments.

Source Data

• Initial Data Collection:

  Data was collected in a utility room with a single transmitter, multiple receivers using ESP32-S3 devices, and a camera (sampling rate ~30 Hz). The setup included:

  • Frequency Band: 2.4 GHz
  • Bandwidth: 20 MHz (52 subcarriers)
  • Protocol: 802.11n
  • Waveform: OFDM
  • Sampling Rate (CSI): ~100 Hz
  • Antenna Configuration: 1 antenna per device
  • Environment: Indoor utility room.

• Who are the source data producers? The data was collected by researchers, with volunteers moving in a controlled utility room environment.

Annotations

• Annotation Process: Raw CSI data is stored in .csv files under directories named by person. Processed data (wiloc.pkl) includes position coordinates and person IDs derived from camera images using scripts in data_process_example. No raw images are provided for privacy reasons.
• Who are the annotators? The dataset creators processed and labeled the data using scripts (process_csi1.py, process_csi2.py, process_cv.py, combine_csi_cv.py) to extract and combine CSI and coordinate data.

Personal and Sensitive Information

The dataset includes person names as directory labels for raw CSI data and person IDs in processed data but does not include raw images or other personally identifiable information beyond CSI patterns and coordinates.

Citation

@article{zhao2024lofi,
  title={LoFi: Vision-Aided Label Generator for Wi-Fi Localization and Tracking},
  author={Zhao, Zijian and Chen, Tingwei and Meng, Fanyi and Cai, Zhijie and Li, Hang and Li, Xiaoyang and Zhu, Guangxu},
  journal={arXiv preprint arXiv:2412.05074},
  year={2024}
}