The combination of technological advances in integrated circuitry, micro-electro-mechanical systems, communication, and energy storage has driven the development of low-cost, low-power sensor nodes. Networking many nodes through radio communication allows for data collection via multihop routing, but the practical limits on available resources and the lack of global control present challenges. Constraints imposed by the limited energy stores on individual nodes require planned use of resources, particularly the radio.
In this dissertation we present Flexible Power Scheduling (FPS), a network scheduling architecture for radio power management specifically tailored towards energy-efficient data gathering and query dissemination in multihop sensor networks. We present empirical results from experiments on Berkeley Motes running two real-world sensor network applications and show that FPS increases end-to-end packet reception and decreases power consumption by 4X over existing power management approaches.