Article Preview
TopIntroduction
Wireless Sensor Networks (WSN) is a concept that first appeared in the military field and has nowadays a number of potential applications in various domains of daily life, such as health care, critical infrastructures or environmental monitoring. The most relevant feature of WSN is the fact that applications are supported by the cooperation of many unattended network nodes wirelessly connected. Due to its nature, network nodes – also known as motes – should have reduced physical dimensions and very low energy consumption featuring, as a consequence, limited processing capabilities. Nevertheless, they must include several types of sensors and actuators, microcontrollers and wireless interfaces so that the same network is able to support a variety of applications. The optimization of the platforms that compose the network is, therefore, a key aspect for the evolution of this concept.
While WSN system software and protocols have been in constant evolution since the debut of this technology in the beginning of the last decade, hardware embedded architectures suited for WSN have not evolved along with their software counterparts as one could expect. WSN platforms are being supported on the same proposed architectures or even by the very same hardware that was presented almost a decade ago. The more representative, most widely used and commercially available ones are the Mica (Hill, 2002) and Telos (Polastre, 2005). This lack of evolution in the WSN nodes' architecture leads to rising difficulties on meeting the strict energy requirements of increasingly demanding applications.
In order to understand and extend the lifetime of WSN, many studies for software estimation of the energy consumed were undertaken (Dunkels, 2007; Hurni, 2011; Hergenroeder, 2010). None of the authors refers the increase of modularity and the adoption of new energy management modules as a solution to the problem. They all focus on software improvements while maintaining the same hardware architecture. Also, the studies are made relying on the sum of isolated figures supplied by the manufacturers of the different devices composing the architecture and not from a full energetic characterization of the involved platforms. A relevant example of how this aspect must not be neglected is described in Trenkamp (2011). Although focused on performance rather than energetic efficiency, it is shown that one must not rely solely on datasheet figures, being important to conduct the complete characterization of the platforms once assembled.
In this article we argue that by improving the hardware embedded architecture of WSN platforms one can increase the network lifetime: directly, through the use of integrated energy management solutions, and indirectly by a modular design.
A more modular platform where the components other than the processing units can be exchanged allows for fast upgrades to new and more efficient solutions. A paradigmatic example would be the adoption of new generation interfaces as the Bluetooth Low Energy’s. In the existing architectures, this would require the redesign and substitution of the installed motes. This is also true when developing a solution for applications with different data rate and power consumption requirements, where the use of several different communication units is an advantage. Good examples of such applications are the Wireless Multimedia Sensor Networks (WMSN), whose demanding requirements cannot be met by the current platforms.