CSI-based passive intrusion detection bound estimation in indoor NLoS scenario

Area of Science:

• Wireless communication
• Signal processing
• Indoor safety systems

Background:

• Passive intrusion detection is crucial for indoor safety.
• Channel State Information (CSI) offers cost-effective, privacy-preserving intrusion detection.
• Existing CSI methods lack detailed analysis of detection bounds.

Purpose of the Study:

• To precisely characterize and model the intrusion detection bound in indoor non-line-of-sight (NLoS) scenarios using CSI.
• To establish the fundamental performance limits of CSI-based intrusion detection.

Main Methods:

• Derived an intrusion-disturbed NLoS channel model considering human impact.
• Developed a mathematical model for the intrusion detection bound based on the channel model.
• Proposed a practical system to estimate the real-world intrusion detection bound.

Main Results:

• Experimental and simulation results validated the derived bound model.
• Demonstrated the effectiveness of the proposed bound estimation system.
• Quantified the farthest intruder position for CSI-based detection in NLoS.

Conclusions:

• The study provides the first precise characterization and model of CSI-based intrusion detection bounds in indoor NLoS.
• The derived model sets a fundamental performance limit for basic CSI intrusion detection.
• Offers a valuable reference for bound estimation in other fine-grained wireless sensing applications.