Drone Detection and Classification Using Physical-Layer Protocol Statistical Fingerprint.

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Drone detection and classification are critical for security and management.
• Existing methods often struggle with accuracy and robustness in complex RF environments.
• Analysis at the data link or higher layers can be insufficient for reliable drone identification.

Purpose of the Study:

• To introduce a novel drone detection and classification method using RF communication link analysis.
• To differentiate WiFi drones from other WiFi devices based on communication characteristics.
• To explore the potential of physical-layer protocol statistical fingerprints for device authentication.

Main Methods:

• Signal detection using Power Spectral Entropy (PSE) to distinguish communication signals from noise.
• Drone classification via physical-layer protocol statistical fingerprint (PLSPF) analysis of extracted communication packets.
• Physical-layer traffic analysis directly from I/Q signals, unlike traditional data link layer methods.

Main Results:

• The proposed method demonstrates scale invariance, frequency invariance, and noise robustness.
• Successfully distinguishes WiFi drones from other WiFi devices by leveraging unique communication requirements.
• Experimental validation confirms the effectiveness and properties of the approach.

Conclusions:

• RF communication link analysis offers a powerful approach for drone detection and classification.
• Physical-layer protocol statistical fingerprinting provides a robust method for identifying drones.
• This technique holds promise for enhancing RF fingerprinting and enabling physical-layer device authentication.