Decision Fusion Mechanism of Oceanic Sensor Networks Based on Fading Acoustic Communication Channel

Automatic target recognition in the atrocious oceanic situation is one of the most difficult areas in oceanic technology. Underwater target recognition based on oceanic sensor networks has been a brand-new research direction. This paper investigates how to use data fusion function of multi-sensor of oceanic sensor networks, and multiple sensors cooperate to fuse different characters taken by radiation noise signals of underwater objectives. In a clustered oceanic sensor networks, sensor members use their observations of the environment to make local binary decisions (H₀/H₁) about whether an event or object has occurred. Local decisions made by local sensor nodes may be lost or corrupted while transmitted to the cluster head. Therefore, each cluster head makes its intermediate decision based decision error ratio and transmission error ratio, and then transmits the intermediate decision to the decision fusion center. Finally, the paper proposes a novel decision fusion mechanism based on Bayesian estimation to obtain optimal decision, where the decision fusion center provides ultimate decision using local decisions of multi-sensor and multi-cluster cooperatively.