Sensor Network Target Detection with Unlabeled Observations
Zachariah Sutton, Peter Willet, Stefano Maranò · 2020
Target detection involves using measurements from sensors to makes a binary decision as to whether or not a target exits at (or near) a particular point in space. In a typical detection scenario, the operator deploys a group of sensors around the tracking space. Measurements from the sensors are sent to a central processing center where the detection decision is made. Typically, each sensor measurement will have other relevant data appended to it, such as a time stamp, the identity of the sensor from which it originated, and the current position of the sensor. This paper uses the term “label” to refer to this supplementary data. In the interest of saving communication bandwidth between the sensors and the processing center, this work will explore the detection performance when the labeling data is omitted.