A Wideband Transient Signal Detection Method Based on Nonlinear Support Vector Machine

In order to solve the problem that the detection performance of transient weak signal is degraded under low SNR, a wideband transient signal detection method based on nonlinear support vector machine (SVM) is proposed. Firstly, based on the support vector machine method, the spectrum of underwater moving target radiated noise is optimally divided in high-dimensional space, and the appropriate Gaussian kernel vector machine is selected to map the input original spectral data to high-dimensional feature space. Then, the similarity or dissimilarity of time-frequency cells (TFC) along the frequency axis is used to detect sudden changes over time. Finally, the transient signals with sudden changes in the frequency spectrum within a short time and wideband range are detected. The data processing results show that the detection ability of the proposed method is improved and the false alarm probability is effectively reduced. By implementing the algorithm on ADSP TS101S, the engineering practicability of the algorithm is proved.