Maximizing Sensing Coverage in Wireless Sensor Networks Through Optimal Scattering of Wake-up Times

Wireless sensor networks (WSNs) often rely on duty-cycling, alternating periods of low-power stand-by with others where sensing, computation, and communication are performed. Duty-cycling brings substantial energy savings, but may complicate the WSN design. The effectiveness of a node in performing its task (e.g., sensing events occurring in an area) is affected by its wake-up schedule. Random schedules lead to deployments that are either ineffective (e.g., insufficient sensing coverage) or inefficient (e.g., areas covered by multiple nodes simultaneously awake). In this paper, we focus on the problem of scattering the nodes' wake-up times optimally, to achieve maximal coverage of a given area. In previous work [5], we presented a decentralized protocol that improves significantly over random wake-up schedules. Instead, here we provide a centralized \emph{optimal} solution that complements the work in [5] by identifying the theoretical upper bound to distributed protocols. Moreover, the modeling framework we present, based on integer programming techniques, is general enough to encompass alternative formulations of the problem. These include the inverse problem of determining the optimal schedule given a desired coverage, as well as other problems based on constraints other than coverage (e.g., latency of data dissemination).