无线传感器网络中以模式识别为基础的事件检测

Pattern Recognition Based Event Detection for Wireless Sensor Networks

实时事件检测是无线传感器网络中的一个重要服务:从连续数据流中识别事件片断并有效的分配系统资源和管理事件数据。作者提出了一个分布式的,实时模式识别机制用以检测无线传感器网络中的敏感事件。该检测方法允许用户进行事件模式查询,网络自动监测用户关注的事件模式。

Real-time event detection is one important service in Wireless Sensor Network （WSN） systems： identifying event segments from continuous data stream to allocate more system resources （radio bandwidth, memory space, etc.） to manage the event data. I propose a distributed, real-time pattern recognition mechanism for event detection in wireless sensor networks. Not like traditional threshold based query, a user is able to query an event pattern, which is a data segment with multiple data samples, into a wireless sensor network. The concerned event is identified after matching real-time stream sensing data with the query pattern. This approach enables the automatic detection of user-concerned event pattern.