부산대학교 도서관

Wireless sensor networks are a current area of interest for many researchers, however surprisingly few have actually been deployed. In this paper we present preliminary results from a series of experiments designed to assess the viability of using existing mobile phone networks in developing countries to create flood warning systems. Creation of a flood warning network involves placing wirelessly connected nodes on all significant tributaries of a river system often over several hundreds of square kilometres. Rain is one of the principal causes of signal change in mobile networks. To assess the effect of the above on associated Internet of Things (IoT) flood alert systems, measurements were conducted in the UK and Mexico to determine the attenuation of line of sight (LOS) and ground multipath propagation due to either rain or excess surface water. Both components are analysed using ray tracing simulation software in addition to the real-time field measurements with the mobile hand set ‘app’ G-NetTrack Pro. Measurement experience gleaned from the UK campaign has been used to formulate a more comprehensive hand set based strategy in the Colima flood zone in Mexico, the results of which are also summarized in this paper. It is believed that both the above measurement and ray trace modelling exercises may be used to further optimize protocols required to address the propagation needs of future 4G/5G networks.