Quality of Monitoring of Stochastic Events by Proportional-Share Mobile Sensor Coverage

There is significant interest in using mobile sensors to protect geographical regions against hazards, in which the sensing resources are distributed according to the varying importance of the sub-regions, such as their numbers of residents exposed to the hazards. The quality of monitoring (QoM) resulting from such proportional-share allocation of the coverage time, in terms of the amount of information captured, is not well understood. In this paper, we analyze the QoM properties of proportional-share mobile sensor coverage, at different fairness time scales, as a function of a wide range of event types, stochastic event staying times, and stochastic event arrival/departure dynamics. Based on the QoM analysis, we optimize a class of periodic mobile coverage schedules that achieve accurate proportional sharing while maximizing the QoM of the total system.