Acoustic Fault Identification of Underwater Vehicles Based on SOM/OMRBF

A network model using Self-organizing map (SOM) and Outputs Modifiable Radial Basis Function (OMRBF) is proposed to identify acoustic fault of underwater vehicles. This model integrates unsupervised SOM with supervised OMRBF to accomplish incremental learning. The outputs neurons of this model can be modified on-line, and SOM is utilized to determine the optimal number of hidden neurons. Experiment results show that the proposed model can identify and remember new faults without forgetting the old ones, and it has good generalization ability.