Rule based inference and action selection based on monitoring data in IoT

Abstract

The current trend in IoT is to find the ultimate solution to integrate objects to the body of Internet to communicate. Once IoT applications are able to incorporate "Things" effortlessly, handling the transferred data is the major challenge. For IoT platforms, when they are mature enough to plug in things with minimal effort, the future research will be around software frameworks. it is fair that IoT in its early years of existence pay much attention to the engagement of things. However, it is predictable that in the future the trend in IoT researches will fall in software area. A typical IoT platform already includes a software framework to handle and manage data. An IoT software framework is a "Rule-Engine" capable of making decisions based on received data. "Expert Systems" has already been on research to address this problem. However, the emergence of IoT will open new doors to this field. Making rule engines for IoT applications differs in that they will process data that is inherently different. In this thesis, an IoT software framework with good level of extensibility is offered which allows developers to easily make IoT solutions on top of that. To analyze data streams, there is an interface to host Machine Learning algorithms, together with other interfaces common for an IoT application. To plug in new extensions, the developer is free to develop their own extensions from scratch or to use some other IoT platforms to integrate new modules.