Advanced feature extraction and dimensionality reduction for unmanned underwater vehicle fault diagnosis

This paper presents a novel approach to the diagnosis of blade faults in an electric thruster motor of unmanned underwater vehicles (UUVs) under stationary operating conditions. The diagnostic approach is based on the use of discrete wavelet transforms (DWT) as a feature extraction tool and a dynamic neural network (DNN) for fault classification. The DNN classifies between healthy and faulty conditions of the trolling motor by analyzing the stator current and vibration signals. To overcome feature redundancy, which affects diagnosis reliability, the Orthogonal Fuzzy Neighbourhood Discriminant Analysis (OFNDA) approach is found to be the most effective. Four faulty conditions were analyzed under laboratory conditions, and the results obtained from experiment demonstrate the effectiveness and reliability of the proposed methodology in classifying the different faults faster and more accurately.