A Real-Time High Sensitivity Signals Detection Method for ELF Wireless Communication

Abstract

To improve real-time monitoring of weak extremely low-frequency (ELF) communication signals in strong-noise environments, this paper proposes a visual detection method that combines short-time Fourier transform (STFT) based time-frequency representation with YOLO-based object detection. With minimum shift keying (MSK) signals as the detection target, a closed-loop experimental system is developed for signal generation, serial-port transmission, real-time spectrogram construction, and automatic visual recognition. At the transmitter, MSK signals are corrupted by adjustable $\boldsymbol{\alpha}$-stable noise and additive white Gaussian noise to simulate complex weak-signal environments. At the receiver, real-time data are converted into time-frequency spectrograms, and weak spectral-line regions are automatically localized using the object detector. Experimental results show that the proposed method can effectively characterize and detect weak MSK signals in real time under low-SNR and impulsive-noise conditions, providing a practical approach to weak ELF signal monitoring and intelligent spectrum sensing.