Energy-efficient coverage with wireless sensors

Abstract

Many sensor networks are deployed for the purpose of covering and monitoring a particular region, and detecting the object of interest in the region. In these applications, coverage is one of the centric problems in sensor networks. Such problem is centered around a basic question: “How well can the sensors observe the physical world?” The concept of coverage can be interpreted as a measure of quality of service provided by the sensing function in various ways depending on sensor devices and applications. On the other hand, sensor nodes are usually battery-powered and subject to limitations based on the available battery energy. It is, therefore, critical to design, deploy and operate a wireless sensor network in an energy-efficient manner, while satisfying the coverage requirement. In order to prolong the lifetime of a sensor network, we explore the notion of connected-k-coverage in sensor networks. It requires the monitored region to be k-covered by a connected component of active sensors, which is less demanding than requiring k-coverage and connectivity among all active sensors simultaneously. We investigate the theoretical foundations about connected-k-coverage and, by using the percolation theorem, we derive the critical conditions for connected-k-coverage for various relations between sensors’ sensing radius and communication range. In addition, we derive an effective lower bound on the probability of connected-k-coverage, and propose a simple randomized scheduling algorithm and select proper operational parameters to prolong the lifetime of a large-scale sensor network. It has been shown that sensors’ collaboration (information fusion) can improve object detection performance and area coverage in sensor networks. The sensor coverage problem in this situation is regarded as information coverage. Based on a probabilistic sensing model, we study the object detection problem and develop a novel on-demand framework (decision fusion-based) for collaborative object detection in wireless sensor networks, where inactive sensors can be triggered by nearby active