Building sensor network surveillance systems: On the applicability.

摘要

A Wireless Sensor Network (WSN) is a self-organized wireless network composed of a large number of sensor nodes deployed in the physical world. The ubiquitous deployment of a huge number of tiny embedded sensors enables a smarter space that better serves people by automatically monitoring and interacting with the physical world. To enable the pervasive usage of WSN systems in practice, an essential issue is to improve the applicability of such distributed, large-scale, remotely working and usually resource constraint systems. Indeed, the sensor network system is highly application driven and its applicability is closely related to the application environment. This dissertation research focuses on developing techniques that makes the sensor network systems more applicable in practical applications. We have been working on a real world project of deploying a sensor network system for underground coal mine surveillance. Throughout the real world project, we address several key problems that emerge from the practice and limit the system applicability, including constructing a structure-aware self-adaptive sensor network framework against the dynamic geological structures in the underground coal mine, designing customized localization and navigation approaches for the WSN systems, introducing a non-threshold based event detection approach that is capable of detecting sophisticated events, and etc. We develop theoretical principles and design practical protocols for implementing our ideas. Through intensive trace-driven simulations and realworld implementations, we evaluate these approaches and the results validate their effectiveness and efficiency. The proposed approaches can be further replanted to other application areas, and we believe widely employing the proposed techniques will significantly improve the applicability of sensor network systems in their practical usage.