Adapting Event Processing in Dynamic IoT Applications, Meeting Evolving Requirements for Quality, Privacy, and Rule Autonomy

The Internet of Things (IoT) connects billions of devices that constantly send information — from sensors in cars and hospitals to smart home systems. Processing this flood of information quickly and responsibly is a major challenge. My PhD research focuses on improving event-based systems, which are the engines that detect important situations from continuous data streams — for example, identifying a fire from combined smoke and temperature readings.
This thesis develops new ways to make these systems more adaptive, private, and autonomous. First, it introduces methods that allow systems to maintain good performance and accuracy, even when sensor quality or availability changes. Second, it proposes techniques to protect users’ privacy, not only by hiding sensitive data but also by preventing the discovery of hidden patterns that could reveal private information. Third, it explores how artificial intelligence, particularly large language models, can automatically create and refine detection rules without human intervention. Finally, it applies these ideas in a realistic scenario involving connected vehicles that share data to improve road safety while preserving driver privacy.
Together, these contributions make event-based processing systems more intelligent, trustworthy, and suitable for the complex and dynamic world of IoT.