Information Granularity With the Self-Emergence Mechanism for Event Detection in WSN-Based Tunnel Health Monitoring

Because of its rapid deployment, low cost, self-organization, node distribution and other characteristics, the wireless sensor networks (WSNs) are very suitable for various types of monitoring systems. Event detection using sensor data is the purpose of the monitoring system. To build the bridge between data and phenomena, this work proposes an information granularity. The information granularity characterizes the health state of each monitoring ring in the tunnel by qualitative description and quantitative representation. Through building the tree augmented naïve Bayesian-based qualitative classifier based on the minimum risk, a qualitative description method of the information granularity is proposed, which is taken as a preliminary judgment of disasters. Inspired by vague set theory, a quantitative representation method of the information granularity is also presented, which analyzes disaster phenomena in detail. To realize the multiscale online monitoring of the entire tunnel state, a kind of self-emergence mechanism of the information granularity is also described. The information granularity is formed gradually through distributed computing during information transmission, which characterizes the event autonomously. The experimental results show that the information granularity realizes rapid, accurate and multiscale representation of tunnel health status.