We established a deep-sea internal wave detection model based on the nonlinear Schrödinger (NLS) equation and Synthetic Aperture Radar (SAR) images, and applied the model to the Malin Shelf edge, located at UK Continental Shelf, west of Scotland, to retrieve internal wave parameters. We selected the SAR images of internal waves at Malin Shelf edge, combined NLS equation with the action spectrum balance equation and Bragg scattering model, retrieved the amplitudes and phase velocities of the internal waves at Malin Shelf edge, and compared these data with those retrieved by the model based on KdV equation and those observed at the same period. The results show that the error between the data retrieved by our model and the measured data is very small, while the difference between the data retrieved by the detection model based on KdV equation and the measured data is significant. In addition, the phase velocities, calculated in our model and the model based on KdV equation, are both close to the measured data. Consequently, our model is valid and more accurate for the parameter inversion of internal waves in deep-sea area.

In order to improve the performance of nonlinear modeling, a Hopfield neural network modeling method based on Subset Kernel Principal Components Analysis (SubKPCA) with Fuzzy C-Means Clustering (FCMC) is proposed. The simulation result shows that, the performance of the proposed method is better than that of Hopfield neural network based on KPCA. It also is effective and feasible to establish the model for the estimation of missing flight data.