Acoustic modulation signal recognition based on endpoint detection.

Area of Science:

• Signal Processing
• Communications Engineering
• Acoustic Signal Analysis

Background:

• Modulation recognition is vital for signal demodulation and communication reconnaissance.
• Acoustic signals (under 20 kHz) present unique challenges for accurate identification.

Purpose of the Study:

• To investigate and enhance modulation recognition technology for acoustic signals.
• To improve signal preprocessing and feature extraction techniques for better accuracy.

Main Methods:

• Compared four endpoint detection algorithms for modulation signal recognition.
• Optimized the short-time energy entropy ratio algorithm with three noise reduction methods.
• Enhanced cyclic spectrum analysis using kurtosis coefficient for differentiating 2FSK and 4FSK signals.

Main Results:

• The optimized algorithm demonstrated effective modulation recognition capabilities.
• Achieved over 99% recognition accuracy for differentiating 2FSK and 4FSK signals at a 4 dB SNR.
• Successful noise reduction and feature extraction were confirmed.

Conclusions:

• The proposed acoustic modulation recognition algorithm significantly improves signal identification accuracy.
• The optimized cyclic spectrum analysis is effective for distinguishing similar frequency shift keying signals.
• This research contributes to advancements in signal demodulation and reconnaissance technologies.