Collaborative Recognition over Distributed Underwater Acoustic Network

Fusion among multi-nodes is an important developing direction of underwater acoustic networks. Information fusion of underwater acoustic (UWA) network should be achieved with minimum energy consumption and aim at achieving a high accuracy of specific application. This paper takes underwater recognition as an example to show how distributed estimation methods can contribute to information fusion over UWA network. The optimization of this process includes individual recognition and recognition of the whole network. In this paper, adjusted diffusion strategy is used to complete network information interaction to make full use of the data from the whole network. Besides, Dempster-Shafer (D-S) model is modified in information interaction process to achieve the decision fusion. This method is verified to be able to accomplish an accurate recognition within the whole network without increasing extra energy cost.