Exploring conflicts in rule-based sensor networks

This paper addresses rule conflicts within wireless sensor networks. The work is situated within psychiatric ambulatory assessment settings where patients are monitored in and around their homes. Detecting behaviours within these settings favours sensor networks, while scalability and resource concerns favour processing data on smart nodes incorporating rule engines. Such monitoring involves personalisation, thereby becoming important to program node rules on the fly. Since rules may originate from distinct sources and change over time, methods are required to maintain rule consistency. Drawing on lessons from Feature Interaction, the paper contributes novel approaches for detecting and resolving rule-conflict across sensor networks.