RainWatch Project: Location-Awared Realtime Detection and Notification of Rain on Internet-Based Sensor Network Kaveevivitchai Sathita Hideya Ochiai Hiroshi Esaki The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan wan@hongo.wide.ad.jp jo2lxq@hongo.wide.ad.jp hiroshi@wide.ad.jp Abstract— This paper introduces RainWatch project which enable rain-sensitive users to monitor real time rain information in area of users’ interest. Real-time detection and notification of rain situation will allow users to take appropriate preparation (e.g., evacuation from the area in case of disaster prevention), which can prevent loss of lives and damage on our properties. The RainWatch system stands on top of Live E! data platform, which is a wide-area sensor application testbed operated by multiple-organizations distributedly with about 100 Internet weather stations. The RainWatch system monitors the rain state around a certain geographical point and makes notifications or warnings (e.g., the rain is approaching) in real-time to the users that have interest in the point. Strong needs for this application will influence the design of the sensor networking infrastructure regard to location-awareness and real-time-ness. I. INTRODUCTION Internet weather stations have enabled realtime collection of environmental data over wide geographical areas. The collected data should be practically managed in a distributed manner because of operational reasons (e.g., multiple-organizations) and of the vast data amount. On the other end, there will be many types of data retrieval requests generated by user applications. To design the system that provides sufficient performance for those requests is one of the major issues in this research area. This paper describes RainWatch Project as one of the applications that provide information in term of users/sensors’ location-aware. This project is aimed to provide useful-information for rain-sensitive end users via e-mail or by other means; e.g., (1) the rain has been observed within 6 km around this point, or (2) the rain is approaching toward this point. This RainWatch service extracts a small set (specified by geographical-area) of raw data from the massive dataset, analyzes and summarizes it into user-friendly format. In other word, RainWatch is the application, which aggregate interested raw data from the location-aware dataset. This application provides the idea in designing Internet-based sensor networks applications. RainWatch project is one of the approaches to tackle on recent natural disasters, which causes lost of lives and destructions; e.g., torrential rain, flash flooding and/or heat island effects in the metropolitan. By observer (end-users) able to receive prior notice about the rain that will occur shortly, they can take appropriate actions in order to minimize those possible effects. We recognize that there are existing methods which provide rain forecast based on Surface Weather Observation, Weather satellites, IR-image, Meteorological radar, together with prediction models [1]. However, this RainWatch project takes advantage of real-time-ness of the system for watching the current rain situation around Point of Interest (POI) and provides the report or notification to the end-users. So they can further take actions in preparation for the coming rain. This output can be used in integral with other methods to provide more accurate notifications or warnings for the end- users. We deploy the RainWatch system over Live E! sensor networking platform [2]. Sensor readings from about 100 Internet weather stations are distributedly managed on the Live E! data platform, and users can retrieve those data for their own activities. By applying the RainWatch system to the real sensor testbed, it is expected that we will get the beneficial examples as follow:  Workers in an outdoor construction site must care about the rain status of their location for their own safety. The workers can halt their work and move out from the field before the rain hits the site.  By previously knowing about the coming rain, stores can pack up products that set outdoor to prevent them from damage  Stores can also enhance some sales such as sale of umbrella by preparing it to be just in grasp for the customers  Disaster prevention in case of flash flooding can be achieved by prior warning II. LIVE E! PROJECT Live E! Project is an open research consortium established by means of collaboration among industry and many academic institutions. This project provides platform for sharing digital environmental data, currently focusing on weather data namely; temperature, humidity, pressure, rainfall, wind speed, and wind direction. These data are collected by sensor nodes embedded with Internet