Adaptive Weighted Envelope Spectrum: A robust spectral quantity for passive acoustic detection of underwater propeller based on spectral coherence

Both surface ships and submarines are vital sea transportation platforms using underwater propellers as propulsion devices. The propeller signal is the primary source of the radiated noise from surface ships and submarines, which can be analysed using demodulation technology, such as the Detection of Envelope Modulation On Noise (DEMON), narrowband demodulation, and cyclostationary analysis. In these methods, the envelope spectrum is the critical spectral quantity for the passive detection of underwater propellers. However, intense and complicated interference noise always hinders the capture of characteristic modulation frequencies in the envelope spectrum. Under the framework of cyclostationary processes theory, the envelope spectrum can be constructed by integrating spectral correlation or spectral coherence along the spectral frequency axis. Selecting a suitable spectral frequency band for integration can significantly enhance the noise immunity of the envelope spectrum and it is also very difficult to implement adaptively. Under this guideline, this paper proposes the Adaptive Weighted Envelope Spectrum (AWES) by allocating weighting coefficients to spectral coherence adaptively based on the modulation intensity of each spectral component. On this basis, the Average Peak-to-threshold Ratio (APR) is constructed for the demodulation capability evaluation of different methods and the automatic detection of characteristic modulation frequencies of propellers. Finally, the effectiveness and superiority of the proposed AWES are validated by simulations and actual signals from water tunnel experiments and merchant ship tests.