Dynamic potential stochastic resonance for weak signal detection

Abstract

Abstract To suppress the low-frequency interference noise, a dynamic potential stochastic resonance (DPSR) model is proposed in this paper for weak signal detection. The DPSR model introduces a single dynamic parameter k that simplifies parameter optimization. Its dynamic potential function can adaptively adjust to match noisy input signals. This model provides a new nonlinear model for triggering the SR phenomenon. Experimental results indicate that compared to the traditional methods that rely on clear interference frequency distributions to filter low-frequency components directly, the SR model offers greater flexibility and convenience. Unlike classical SR models, the proposed DPSR model demonstrates a 1.5 dB improvement in output performance for suppressing low-frequency interference. Therefore, the DPSR model not only robustly suppresses interference but also effectively enhances and detects characteristic signals in variant-noise environments. Application to sea trial signals highlights the superior performance of the DPSR model in significantly reducing low-frequency interference and improving target signal recognizability compared to other models.