Participatory sensing is a cost-effective and efficient way to monitor distributed events-based processes, for example, mobile phone usage monitoring. In such processes, two levels of detection or measurement are required: detect an event through coarse measurements, and then measure the process precisely for the duration of the event.

The authors present a cost-effective heuristic solution through two-tier quality of information (QoI) specifications. In their event-driven participatory sensing model, event boundaries for multiple concurrent tasks are approximated by participatory sensing first, followed by the deployment of a set of sensors selected through an optimizing strategy ensuring higher QoI, as required.

In this comprehensive, well-written paper, the authors discuss related works, besides presenting a formal system model; an event boundary detection algorithm, both for coarse-grained boundary detection and fine-grained boundary detection; and experimental results. The paper discusses in some detail the QoI satisfaction metric, quantifying the expected data volume using a mobility model, and energy consumption.

The experimental results, based on a real trace for mobile phone usage in Beijing, show that the proposed methods could be quite effective and robust. This paper should interest both practicing and research communications engineers. It lists nearly 50 references.