Robust Event Boundary Detection in Sensor Networks - A Mixture Model Based Approach

Abstract

Detecting event frontline or boundary sensors in a complex sensor network environment is one of the critical problems for sensor network applications. In this paper, we propose a novel algorithm for event frontline sensor detection based on statistical mixture methods with model selection (Akaike, 1973). A boundary sensor is considered as being associated with a multimodal local neighborhood of (univariate or multivariate) sensing readings, and each non-boundary sensor is treated as being with a unimodal sensor reading neighborhood. Furthermore, the set of sensor readings within each sensor's spatial neighborhood is formulated using Gaussian mixture model (McLachlan and Peel, 2000). Two classes of boundary and non-boundary sensors can be effectively classified using the model selection techniques for finite mixture models. Our extensive experimental results demonstrate that our algorithm effectively detects the event boundary with a high accuracy under moderate noise levels.