Efficient target detection in maritime search and rescue wireless sensor network using data fusion

Maritime search and rescue wireless sensor network (MSR-WSN) has been a bedrock to discover the floating target after the shipwreck. In this paper, we first define a sea region of target detection and formulate a clustered topology of MSR-WSN. Second, we employ the sensor nodes of MSR-WSN to track the collective radio signal emitted by the mobile target. Each node firstly transmits the preprocessed perceived data to the cluster head node. Next, the data fusion center (DFC) collects a local decision of cluster head node through a binary hypothesis test and works out an accurate global decision. This paper emphasizes at designing both local and global data fusion rules based on the likelihood of ratio test statistics using a Neyman-Pearson lemma and Bayesian approach. One major stumbling block in the ocean lies in a complex and changing communication environment. There is a need for the DFC to develop a fusion rule of carrying out a dependable target detection to screen out the side effect of wave shadow. To address the concern, we propose a novel mobile target detection algorithm (NMTDA) based on information theory. The main idea is to dynamically calculate an adaptive decision threshold using both Kullback-Leibler divergence (KLD) and a global optimal decision statistics to enforce the accuracy of target detection. In addition, KLD is adopted to quantify the strength of wave shadow effect and tune Correct Detection/Flase Alarm probabilities of target detection. To conserve the overall MSR-WSN energy, DFC selects clusters with the maximum predictive information gain for MSR before next round search. Extensive simulation results demonstrate that our proposed mobile target detection algorithm works well in maritime search and rescue scenario.