Hybrid Intelligent Detection for Underwater Acoustic Target Using EMD, Feature Distance Evaluation Technique and FSVDD

In order to solve the problem of accurately detecting the weak acoustic signal for remote underwater target, a novel hybrid intelligent target-detection method for underwater acoustic signals based on empirical mode decomposition (EMD), feature distance evaluation technique (FDET) and fuzzy support vector data description (FSVDD) is proposed. The method consists of three stages. Firstly, some signal processing methods, like filtration, Hilbert envelope-demodulation and EMD are carried out to extract the time- and frequency-domain statistical features from original underwater acoustic signals, and these features make up an integrated feature set. Secondly, with the FDET, the salient feature set is obtained from the integrated feature set. Finally, the salient features are input into the detector based on FSVDD to detect the underwater targets intelligently. This method is applied to target detection of underwater vehicle. Testing results show that the proposed method has better detection performance than the traditional detector based on SVDD, with a high detection success rate.