Wireless Sensor Networks (WSNs) enable new applications that involve a tight couplingbetween conventional computing infrastructure and the physical world. While the potential ofsensor networks is only beginning to be realized,several challenges still remain.
Arguably, theforemost among these is satisfying the requirementfor long-lived operation (months to years)for several sensor network applications. Due to the limited capacity of batteries and the difficulty of frequent battery recharging or replacement, energy is a scarce and precious resource in sensor networks.
Decreasing the energy consumed during system operation, which directly translates toincreased lifetime, has been a goal of much sensor network research in hardware and software design, network protocols, and middleware services.
Despite the considerable research attention that energy optimizationhas received, the problem of network lifetime as yet remains unsolved. As sensor networks make the transition from synthetic labscaletest-beds to real-world deployments, it is more important than ever to identify promising techniques that can yield significant energy benefits.
This article discusses several such emerging approaches that we believe will help achieve the goal of long-lasting sensor networks. Also described is a new, energy-efficient sensor-node architecture using a a heterogeneous multiprocessorwith staged wakeup.
It uses two types of processors.In the “low workload” phase the high-end processor is power gated (i.e., completely powered off) and the sensor node uses allow-end processor that provides low overhead wakeup.
When the sensors detect activity, the network transitions to the “high workload”phase. The high-end processor awakened and performs the intensive computation and communication in an energy-efficient manner.
Once the intruder leaves, the network transitions to the “low workload” phase again and the high-end processor is shut down.The article also includes a discussion of ultra-low power medium access protocols as well as a description of techniques for environmental energy harvesting and for energy optimization of the sensing subsystem.
To read this external content in full, download the complete paper from the author online archives at the University of California, Los Angeles.