A Generic State Model with Neighbourhood Support from Wireless Sensor Networks for Emergency Event Detection
Wireless Sensor Networks typically consist of a large number of sensor nodes deployed in a physical space, for the purpose of detecting the existence of interesting patterns or events. A variety of event detection approaches have been researched, however, they are limited to the specific event detection scenarios. Based on what constitutes normal patterns of behaviour, the existing event detection approaches can be classified into two categories: threshold based event detection and tempo-spatial pattern based event detection. In spite of the differences in the implementation of event detection algorithms, the common problem of event detection is the reliability issue. According to the latest UK fire statistics published (1997-2007), every year more than half of the fire-related incidents attended by fire and rescue services were false alarms. The result of these was a waste of resources responding to nonexistent emergencies, and a risk of not having enough resources available for real fire emergencies occurring soon after. A generic state model with Neighbourhood Support from Wireless Sensor Networks for emergency event detection is proposed, which addresses the reliability issue and takes the features of sensor data into account. The model operates in an event-driven mode, in contrast to constant monitoring. The detection process involves 1) sensor node level state model for suspicious behaviour detection, 2) Neighbourhood Support checking to confirm or deny the event, and 3) network level fusion to characterize the event (e.g. the affected area, the type and the severity of the event). The concept of Neighbourhood Support is proposed to separate real alarms from outliers, thus improving the reliability of emergency event detection. The application of the generic model to both threshold based event detection and tempospatial pattern based event detection are discussed. The simulation results demonstrate that the proposed state model with Neighbourhood Support can apply in both event detection categories and can improve the reliability and efficiency of the existing event detection approaches.
emergency event detection sensor state model neighbourhood support Wireless Sensor Networks emergency response
YANG Yanning Andrew J.MAY YANG ShuangHua
Computer Science Department, Loughborough University, Loughborough LE11 3TU, UK Ergonomics and Safety Research Institute, Loughborough University, Loughborough LE11 3TU, UK
国际会议
The 2010 International Symposium on Safety Science and Technology(2010 安全科学与技术国际会议)
杭州
英文
365-376
2010-10-26（万方平台首次上网日期，不代表论文的发表时间）