Agent Oriented End-User Programming for the Web of Things

In the context of a Pan-European project aimed at defining the next generation of intelligent IoT systems, this thesis proposes the idea of Visual Agent-Oriented Programming as the enabling tool for non-technical users to configure and program such systems. The thesis work was carried out while being hosted by the University of St. Gallen, contributing to the exploration of the Interaction and Communication-based Systems research group in the field of engineering autonomous systems capable of controlling Web of Things (WoT) environments with a human-in-the-loop philosophy. Motivated by the ever-increasing demand for interfaces designed for workers to be able to keep the pace of the fast digitalization of business activities, a new tool was designed and developed to provide a seamless interface to control the digital representation of physical environments exploiting agent programming. To our knowledge, this is also the first attempt at building a user-friendly interface on top of the agent-oriented paradigm that targets people with no previous programming experience. This can be seen as the first step in exploring whether the assumptions made when defining agents as entities to build software in a way that was more understandable for humans since taking inspiration from models of human behaviour and reasoning still hold. From a WoT perspective, providing a uniform programming interface is crucial to allow to create or modify software featuring a different degree of autonomy in flexibly performing tasks, dealing with open, dynamic and distributed environments. Users can leverage the agent paradigm to define complex behaviour at a very high level of abstraction leaving space for all sorts of reconfiguration to happen in the lower layers such as on-demand machine allocation.