17th Telecommunications forum TELFOR 2009 Serbia, Belgrade, November 24-26, 2009. Abstract – Wireless Sensor Networks (WSNs) are rapidly emerging as a potential networking technology for usage in various emergency situations. WSNs offer capability to locally and collaboratively sense and interpret collected data from the environment and, correspondingly, react to various situations. This paper presents a small scale WSN testbed for early fire detection. The testbed comprises wireless sensor nodes – Sun SPOTs, different additional sensors and appropriate accessories which provide real time monitoring of the environmental parameters and on-time notification system. The testbed serves as a proof-of-concept and is further extendable for many different applications and outdoor settings. Key words – Fire, Sensors, Sun SPOT, Testbed, WSN. I. INTRODUCTION HE world witnesses disasters of various magnitudes caused by global warming, different social and political movements, natural catastrophes etc. These phenomenons, if uncontrolled, can cause significant damages to natural and human resources. For example, in August 2003, a forest fire was initiated by lightning a strike in the Okanagan Mountain Park in the Province of British Columbia, Canada. The estimation showed that 25 912 hectares were affected by the fire and 239 homes were burned [1]. Furthermore, in Macedonia, during the summer 2007, around 400 fires affected 70 000 ha of forest causing €32 million material losses. Fires also have serious economic implications such as destruction of habitats, forest and material damage, costs of fire-fighting etc. Because of these destruction issues, prevention and efficient monitoring of fires has become a global concern. Wireless Sensor Networks (WSNs) lately provide means to observe hazardous phenomenons. They represent wireless networks of spatially distributed sensor nodes used for monitoring different physical phenomena and parameters and transferring the data in real time to the data acquisition center for additional processing. Owing to these characteristics, WSNs are emerging as a novel and efficient concept for fires forecasting and detecting. They constitute a promising framework for building near real- time fire detection systems. This paper presents a demo platform, as a proof-of- concept, for WSN based early fire detection system. The testbed consists of wireless sensor modules – Sun SPOTs, All authors are with the Faculty of Electrical Engineering and Information Technologies, Ss. Cyril and Methodius University of Skopje, Macedonia (phone: 389-2-3099114; fax: 389-2-3064262; e-mail: liljana @feit.ukim.edu.mk) . different sensors, a server and a GSM/GPRS gateway. The idea of the demo realization is to provide remote real time monitoring of environmental parameters of interest, detect possible fire and send fire alarm reports to the responsible departments. The rest of the paper is organized as follows. Section II describes the scenario for early fire detection, the basic idea and the developed algorithms. Section III analyzes the performed experiments. Finally, section IV concludes the paper. II. TESTBED FOR EARLY FIRE DETECTION This section elaborates on a small scale demo setup for early fire detection using WSN. A. Scenario Description The scenario setup consists of Sun SPOTs [6], four types of sensors and a base station forming a WSN, MySQL database hosted on a server, GPS/GPRS gateway and web application - GUI. The demo platform utilizes 3 Sun SPOTs, 3 temperature/humidity sensors, 3 carbon dioxide and 1 wind speed sensor. These sensors can measure the key parameters which indicate fire ocuurance and the possibility of fire spreading. Fig. 1. Relation between humidity and temperature obtained with daily measurements The measurements of the daily temperature and humidity variations in Fig. 1 show that the time interval when a fire can occur (when the value of the temperature is higher than the humidity) is between 12 and 19 hour. In other words, when the weather is dry and hot, a fire can occur with higher probability. This information is used as an input value for the development of corresponding algorithms for fire detection in the testbed. Fig. 2 depicts the scenario and the testbed architecture which is further elaborated in details. Early Fire Detection with WSN Sanja Bonevska, Andrijana Jankova, Aleksandra Mateska, Vladimir Atanasovski, Student Member, IEEE, Liljana Gavrilovska, Member, IEEE T 1482