Mathematical Analysis of Traffic Detection Systems Using Unreliable Sensors.
An initial attempt to deal with a new class of applied mathematical problems: how to extract information optimally from emplaced fields of sensors. These problems are relevant to interdiction and border security. The basic sensor model assumes that each sensor is subject to both false alarms and detection failures. The analytical assumptions contained in this statistical model represent a compromise between the opposing demands of realism and mathematical tractability. Methods are presented for determining the properties of traffic flow that can be derived from the data produced by one such unreliable sensor, and by two or more of them. A method is given for formulating a suboptimal strategy for deploying a fixed number of perfectly reliable sensors to monitor a territorial border. An understanding of probability theory is assumed; detailed derivations are included as appendixes. 53 pp.