Energy-Aware Selective Communications in Sensor Networks
Rocío Arroyo-Valles, Gianluca Rizzo Antonio, Jesús Cid‐Sueiro · InTech eBooks · 2011
During the last years, Wireless Sensor Networks (WSN) have attracted the attention of researchers from electronics, signal processing, communications, and networking communities due to their potential for providing new capabilities. Among the many design challenges that have been identified, the ability of sensors to behave in an autonomous and self-organized manner using limited energy and computation resources has emerged as a fundamental factor to take into account when WSN are deployed. In fact, the limitation of resources at the network nodes is often a critical factor that conditions the design of applications for sensor networks. Among the multiple limitations to consider, energy consumption emerges as a primary concern. This is because in many practical scenarios, sensor node batteries cannot be (easily) refilled, thus nodes have a finite lifetime. Since every task carried out by the WSN has an impact in terms of energy consumption, an enormous variety of solutions, both software and hardware, have been proposed in the literature to optimize energy management; see, e.g., (Shih et al., 2001; Akyildiz et al., 2002). Communication processes are typically among the most energy-expensive of such tasks. Many works have focused on the minimization of the energy cost taking into account the physical behavior of the WSN; see, e.g., (Shih et al., 2001; Marques et al., 2008; Wang et al., 2008). However, energy savings can also be obtained by taking a higher level approach and considering the different nature of the information that nodes have to transmit. This way, in order to enlarge the network lifetime and optimize the overall network performance, sensor nodes should weigh up: (a) the potential benefits of transmitting information and (b) the cost of the subsequent communication process. A first step to address such optimum design is to properly quantify or estimate both costs and benefits. This is possible in many practical cases because the energy consumed during the different communications tasks (cost) is typically well-characterized and because applications where messages are graded according to an importance indicator (benefit) are frequent in WSN. The message importance can be, for instance, a priority value established by the routing protocol, or an information value specified by the application supported by the sensor network. Relevant examples in the context of Sensor Networks can be found in the fields of: security (attack reports (Wood & Stankovic, 2002)), medical care (critical alerts (Shnayder et al., 2005)), or data fusion (DAIDA algorithm in (Qiu et al., 2005)), to name a few. In such scenarios, energy in WSN can be saved by making intelligent importance-driven decisions about message transmission, in an autonomous and self-organized manner, adapting 8