Distributed detection with a scan statistic: Global to local inference
Xiufeng Song, Peter Willett, Joseph Glaz, Shengli Zhou · 2012
Sensor networks offer an opportunity for improving submarine detection. Let each sensor firstly make a binary local decision-`0' or `1', and then a fusion center collects them to make a system-level inference without knowledge of probabilities of local detection. An obvious strategy is a counting rule test, which simply counts the total number of 1's and compares it to a threshold. This approach equally considers all network subareas. However, reflected signals from a submarine are highly aspect dependent, and in many instances only sensors in a particular zone could receive the echoes. This paper focuses on the scan statistic, which slides a window across the sensor field, and selects the subarea with the largest number of 1's to make a decision. The scan statistic integrates the spatial distinction of local decisions into detection fusion. With a proper window size, it may suppress subarea interference, and improve system-level performance.