Natural Disaster Monitoring with Wireless Sensor Networks: A Case Study of Data-intensive Applications upon Low-Cost Scalable Systems

摘要

The wireless sensor network (WSN) technology has applied in monitoring natural disasters for more than one decade. Disasters can be closely monitored by augmenting a variety of sensors, and WSN has merits in (1) low cost, (2) quick response, and (3) salability and flexibility. Natural disaster monitoring with WSN is a well-known data intensive application for the high bandwidth requirements and stringent delay constraints. It manifests a typical paradigm of data-intensive application upon low-cost scalable system. In this study, we first assessed representative works in this area by classifying those in the domains of application of WSNs for disasters and optimization technologies significantly distinguishing these from general-purpose WSNs. We then described the design of an early warning system for geohazards in reservoir region, which relies on the WSN technology inspired by the existing work with focuses on issues of (1) supporting reliable data transmission, (2) handling huge data of heterogeneous sources and types, and (3) minimizing energy consumption. This study proposes a dynamic routing protocol, a method for network recovery, and a method for managing mobile nodes to enable real-time and reliable data transmission. The system incorporates data fusion and reconstruction approaches to bring together all data into a single view of the geohazard under monitoring. A distributed algorithm for joint optimal control of power and rate has been developed, which can improve utility of network (> 95 %) and to minimize the energy consumption (reduction by > 20 % in comparison with LEACH). Experimental results indicate the potentials of the proposed approaches in terms of adapting to the needs of early warning on geohazards.