Power-Efficient Adaptable Wireless Sensor Networks

Wireless sensor networks represent a new data collection paradigm in which adaptability plays an important role. Typical sensor network scenarios involve scattering a large number of wireless nodes from an aircraft across an area of interest. The nodes then form a network through which collected data is routed to a base station. Adaptation is necessary to deal with the unpredictable network topologies that result from sensor node scatters and to manage resources (energy in particular) efficiently in response to changing conditions and requirements. The hardware flexibility required for dynamic adaptation is traditionally achieved with softwarebased processors or field programmable gate arrays (FPGAs), both of which come with significant energy, area and performance costs when compared to application-specific integrated circuits (ASICs). We therefore introduce a smallscale reconfigurable design technique that minimizes these costs by efficiently integrating small amounts of applicationspecific reconfigurable logic within primarily fixed-logic circuitry. This technique provides the flexibility necessary for the adaptations required of wireless sensor networks without the penalties associated with processors and FPGAs. This paper makes the case for small-scale reconfigurability by investigating several different types of adaptation in wireless sensor network applications that allow applications to deal with unpredictable network topologies and tradeoff between network longevity and fidelity, security and latency.