The vision of ubiquitous computing requires the development of devices and technologies that can be pervasive without being intrusive. The basic component of such a smart environment will be a small node with sensing and wireless communications capabilities, able to organize itself flexibly into a network for data collection and delivery. Building such a sensor network presents many significant challenges, especially at the architectural, protocol, and operating system level. Although sensor nodes might be equipped with a power supply or energy scavenging means and an embedded processor that makes them autonomous and self-aware, their functionality and capabilities will be very limited. Therefore, collaboration between nodes is essential to deliver smart services in a ubiquitous setting. New algorithms for networking and distributed collaboration need to be developed. These algorithms will be the key for building self-organizing and collaborative sensor networks that show emergent behavior and can operate in a challenging environment where nodes move, fail, and energy is a scarce resource.
The question that rises is how to organize the internal software and hardware components in a manner thatwill allowthem towork properly and be able to adapt dynamically to new environments, requirements, and applications. At the same time the solution should be general enough to be suited for as many applications as possible. Architecture definition also includes, at the higher level, a global view of the whole network. The topology, placement of base stations, beacons, etc. is also of interest. In this chapter, we will present and analyze some of the characteristics of the architectures for wireless sensor networks. Then, we will propose a new dataflow-based architecture that allows, as a new feature, the dynamic reconfiguration of the sensor nodes software at runtime.
|Title of host publication||Embedded Systems Handbook|
|Place of Publication||Boca Raton, FL, USA|
|Publication status||Published - 2006|
- CAES-PS: Pervasive Systems