CNN-Based Hybrid Optimization for Anomaly Detection of Rudder System

In this study, an automatic test platform suitable for steering gears was established, which can four sets of rudder systems separately. addition, we propose anomaly detection method based on deep learning technology to complete the automated multi-fault classification gear test. This paper combines particle swarm optimization algorithm and grey wolf optimize convolutional neural networks (HPSOGWO-CNN). The proposed HPSOGWO-CNN model is constructed in two stages realize efficient high-accuracy system. first stage, through 10-fold cross validation, optimal number search agents HPSOGWO obtained, performance compared with GWO PSO algorithms respectively. results demonstrate that excellent technique selection hyper-parameters. second designed used fine-tune hyper-parameters CNN, a highly matched system parameters finally obtained. experimental show accuracy 99.846%, precision 99.748%, recall 99.498%, F-score 99.618%, Kappa reaches 0.99565. CNN-based hybrid system, advanced comparison KNN, SVM, BP, PSO-CNN, GWO-CNN, MGWO-CNN, WdGWO-CNN, RW-GWO-CNN models, terms accuracy, precision, recall, F-score, kappa, Moreover, it not affected by imbalance samples, achieve accurate small training samples.