Practical and Robust Power Management for Wireless Sensor Networks

Wireless Sensor Networks: WSNs) consist of tens or hundreds of small, inexpensive computers equipped with sensors and wireless communication capabilities. Because WSNs can be deployed without fixed infrastructure, they promise to enable sensing applications in environments where installing such infrastructure is not feasible. However, the lack of fixed infrastructure also presents a key challenge for application developers: sensor nodes must often operate for months or years at a time from fixed or limited energy sources. The focus of this dissertation is on reusable power management techniques designed to facilitate sensor network developers in achieving their systems' required lifetimes. Broadly speaking, power management techniques fall into two categories. Many power management protocols developed within the WSN community target specific hardware subsystems in isolation, such as sensor or radio hardware. The first part of this dissertation describes the Adaptive and Robust Topology control protocol: ART), a representative hardware-specific technique for conserving energy used by packet transmissions. In addition to these single-subsystem approaches, many applications can benefit greatly from holistic power management techniques that jointly consider the sensing, computation, and communication costs of potential application configurations. The second part of this dissertation extends this holistic power management approach to two families of structural health monitoring applications. By applying a partially-decentralized architecture, the cost of collecting vibration data for analysis at a centralized base station is greatly reduced. Finally, the last part of this dissertation discusses work toward a system for clinical early warning and intervention. The feasibility of this approach is demonstrated through preliminary study of an early warning component based on historical clinical data. An ongoing clinical trial of a real-time monitoring component also provides important guidelines for future clinical deployments based on WSNs.

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science and Engineering

Year dc:date.available

2011

Author dc:creator

• Hackmann, Gregory

Contributors dc:contributor

• Chenyang Lu

• Computer science

Language dc:language

English (en)

DOI

https://doi.org/10.7936/K7TT4P0R

OAI identifier oai:identifier

oai:openscholarship.wustl.edu:etd-1582