It is useful to protect residents in a geographical region against hazards and attacks using a sensor network. Real world sensors for harmful plumes have limited ranges. If there is insufficient sensing resource to cover whole geographical regions all the time, mobile sensor coverage is useful. In this case, we would like to allocate the sensing resources according to the different threat levels of the subregions (e.g., to give more coverage to more densely populated subregions). In this talk, I will analyze the information capture of such threat-based mobile coverage in terms of stochastic event arrivals and departures and the types of event. Based on the analysis, I will present the optimization of mobile coverage schedules such that threat-based coverage is achieved while the information capture is maximized.
Colloquium
