Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

FreeSense:Indoor Human Identification with WiFi Signals (1608.03430v1)

Published 11 Aug 2016 in cs.HC

Abstract: Human identification plays an important role in human-computer interaction. There have been numerous methods proposed for human identification (e.g., face recognition, gait recognition, fingerprint identification, etc.). While these methods could be very useful under different conditions, they also suffer from certain shortcomings (e.g., user privacy, sensing coverage range). In this paper, we propose a novel approach for human identification, which leverages WIFI signals to enable non-intrusive human identification in domestic environments. It is based on the observation that each person has specific influence patterns to the surrounding WIFI signal while moving indoors, regarding their body shape characteristics and motion patterns. The influence can be captured by the Channel State Information (CSI) time series of WIFI. Specifically, a combination of Principal Component Analysis (PCA), Discrete Wavelet Transform (DWT) and Dynamic Time Warping (DTW) techniques is used for CSI waveform-based human identification. We implemented the system in a 6m*5m smart home environment and recruited 9 users for data collection and evaluation. Experimental results indicate that the identification accuracy is about 88.9% to 94.5% when the candidate user set changes from 6 to 2, showing that the proposed human identification method is effective in domestic environments.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Tong Xin (5 papers)
  2. Bin Guo (150 papers)
  3. Zhu Wang (72 papers)
  4. Mingyang Li (86 papers)
  5. Zhiwen Yu (78 papers)
Citations (116)

Summary

An Overview of FreeSense: Indoor Human Identification Utilizing WiFi Signals

The paper "FreeSense: Indoor Human Identification with WiFi Signals" by Tong Xin et al. proposes an innovative approach for non-intrusive human identification using WiFi signals, capitalizing on the Channel State Information (CSI) as a medium to detect unique human movement patterns. This method offers a privacy-preserving alternative to traditional identification systems such as fingerprint and iris recognition, which require close proximity to sensing devices.

Methodological Framework

The FreeSense system operates by leveraging inherent variations in WiFi signals caused by individual body shape differences and motion patterns. The authors advance the technique by focusing on specific events where the human subject crosses a line-of-sight (LOS) path, which results in distinct CSI fluctuations. These fluctuations are isolated and analyzed to identify individuals accurately.

  1. Identification Condition and CSI Time Series Segmentation:
    • The approach determines functional windows by identifying LOS path crossings, using the characteristic CSI waveform as the basis.
    • An algorithm iteratively segments the CSI time series to extract these LOS waveforms.
  2. Feature Extraction and Model Design:
    • The primary method for feature extraction uses a combination of Principal Component Analysis (PCA) to reduce data dimensionality and Discrete Wavelet Transform (DWT) to extract salient features of the waveform.
    • Dynamic Time Warping (DTW) is employed to align waveforms, accommodating different lengths and misalignments in the segmented signal data.
  3. Classification:
    • A k-nearest neighbor (KNN) classifier identifies individuals based on the extracted waveform features.

Experimental Setup and Results

The system was deployed in a 6x5m smart home environment, with experiments involving nine participants. Using WiFi equipment comprising a Lenovo X200 laptop with an Intel Link 5300 WiFi NIC and a TP-Link TLWR1043ND router, CSI data was collected at a rate of approximately 1000 packets per second.

  • Detection and Accuracy:
    • The segmentation algorithm achieved a detection ratio of 92.6% and demonstrated a lower error ratio (6.5%) compared to the baseline method.
    • Classification accuracy varied between 88.9% and 94.5% as the number of users in the test set ranged from 2 to 6, showcasing the system's effectiveness in small domestic settings.
  • Performance under Varied Conditions:
    • When evaluating more extensive walking conditions and varying training set sizes, the system maintained acceptable performance, with recognition accuracy improving from 75.0% to 91.7% as training samples increased.

Implications and Future Directions

This research underscores the potential of WiFi-based identification as a low-cost, pervasive alternative that minimizes privacy invasions, a significant limitation of vision-based systems. Its practicality is evident in smart home applications where personal identification can enhance personalized service delivery.

The findings also open avenues for further research in the field. Future investigations may address factors affecting performance, such as multipath interference, varying environmental contexts, and presence of multiple subjects. Exploration into more complex machine learning models beyond KNN might yield even higher accuracy and robustness.

Overall, FreeSense exemplifies how leveraging ambient infrastructure can lead to innovative approaches in real-world biometric applications, indicating substantial promise for future developments in ubiquitous and context-aware computing environments.

Youtube Logo Streamline Icon: https://streamlinehq.com