Intelligent home automation security system based on novel logical sensing and behaviour prediction

Abstract

The thesis, Intelligent Home Automation Security System Based on Novel Logical Sensing and Behavior Prediction, was designed to enhance authentication, authorization and security in smart home devices and services. The work proposes a three prong defensive strategy each of which are analyzed and evaluated separately to drastically improve security. The Device Fingerprinting techniques proposed, not only improves the existing approaches but also identifies the physical device accessing the home cybernetic and mechatronic systems using device specific and browser specific parameters. The Logical Sensing process analyses home inhabitant actions from a logical stand point and develops sophisticated and novel sensing techniques to identify intrusion attempts to a home’s physical and cyber space. Novel Behavior prediction methodology utilizes Bayesian networks to learn normal user behavior which is later compared to distinguish and identify suspicious user behaviors in the home in a timely manner. The logical sensing, behavior prediction and device fingerprinting techniques proposed were successfully tested, evaluated and verified in an actual home cyber physical system. The algorithms and techniques proposed in the thesis can be easily modified and adapted into many practical applications in Industrial Internet of Things, Industry 4.0 and cyber-physical systems.