Seafloor backscatter signal simulation and classification

Abstract

In this paper, simulation study is initiated to observe the interaction effect of the sound signal with different seafloors types and its classification. Simulation of the echo envelope signal is done by accurately formulating a mathematical model. In this model a smooth echo envelope is generated then mixed up with multiplicative and additive noise. Several such echo signals were simulated for three types of seafloor. An Artificial Neural Network based classification technique is conceived to classify these echo signals based on the various features like peak amplitude, energy and fractal dimension. The seafloor classifications realized using ANN's Multi-layer Perceptron, efficiently performs the task with high classification percentages