Underwater acoustic target recognition algorithm based on EK-NN

In underwater acoustic target recognition, the acoustic signal of the target is usually complex and also has some uncertain information. In order to effectively solve these problems, a new underwater acoustic target recognition algorithm based on evidence k-nearest neighbor (EK-NN) theory is presented in this paper. In this new method, the basic belief assignments (bba) are determined by using the feature of distance between the object and its K-nearest neighbors in each class of the training set, and then the bba in each class are combined with Dempster-Shafer's (D-S) rule. Finally the combined results in each class are fused with Redistribute conflicting mass proportionally rule5 (PCR5), thus the object can be recognized by the above fusion result and the classification rule presented in this paper. Several experiments based on the underwater acoustic data sets were performed to verify the effectiveness of EK-NN in comparison with other methods. The experimental results indicate that EK-NN can effectively improve the recognition accuracy.