R-DEHM: CSI-Based Robust Duration Estimation of Human Motion with WiFi

Area of Science:

• Wireless Sensing
• Human Behavior Recognition
• Signal Processing

Background:

• Human motion duration is crucial for behavior recognition but current wireless methods lack duration estimation.
• Existing wireless sensing technologies primarily focus on presence detection, not duration.

Purpose of the Study:

• To propose a novel system for robust duration estimation of human motion (R-DEHM) using WiFi signals.
• To address the gap in estimating human motion duration leveraging wireless signals.

Main Methods:

• Collecting channel state information (CSI) from commodity WiFi devices.
• Extracting robust features from CSI amplitude.
• Utilizing a back propagation neural network (BPNN) for human motion detection by segmenting CSI sequences.

Main Results:

• Achieved an average human motion detection rate exceeding 94%.
• Demonstrated an average error rate for duration estimation below 8%.
• Successfully transformed continuous duration estimation into discrete motion detection.

Conclusions:

• The proposed R-DEHM system offers a robust and accurate method for human motion duration estimation using WiFi.
• This approach enhances wireless behavior recognition capabilities by incorporating motion duration.
• The system demonstrates practical feasibility and effectiveness in real-world applications.