Center for Embedded Network Sensing

We propose an on-line fault detection and diagnosis framework in sensor networks, called Inspect, which is based on a hybrid tiered approach to integrity checking. Inspect tiered design, combines the benefits of distributed and centralized approaches, thus improving the responsiveness and efficiency of the fault detection. More precisely, Inspect consists of a local tier built at each sensor node and capable of detecting anomalies, and a centralized tier built at the sink and capable of distinguishing between sensor faults and unexpected temporal-spatial variations in the phenomenon, and detecting more complex types of faults.

Inspects offers several desirable features: it provides online detection, it is not designed for a specific application or statistical distribution of the phenomenon, and it is more robust to temporal-spatial variations of the phenomenon. Moreover, Inspect provides confidence bounds that can be dynamically tuned according to the user requirements for a better resource management, and addresses the problem of detecting more complex types of faults, which is a challenging and yet unexplored problem.