The interplay between signal processing and networking in sensor networks

This paper provides a signal processing perspective on large-scale sensor networks. It focuses on two characteristics of sensor networks: application specificity and energy constraint. The former requires that network protocols be designed for the application at hand, which is often signal processing in nature, and measured by application-specific metrics. The latter calls for novel distributed signal processing techniques to provide accurate and timely network state information that can be exploited by network protocols to ensure energy efficiency. The underlying theme is about how a principled integration of signal processing and networking can lead to an efficient and fair use of limited resources. The paper aims to demonstrate that capturing and exploiting dependencies between signal processing and networking offer design choices resulting in improved network performance.