A smart IoT middleware for comfort management based on Multi-Agent-System
Benjamin Gâteau · 2018
In this paper we propose a new approach of comfort management based on a smart IoT middleware composed of a Knowledge Base (KB) and a Multi-Agent System (MAS). Those elements allow users in a room to interact with the system to request in a simple and natural way the change of one or more comfort elements (including temperature, humidity, luminosity, noise, ventilation and air quality). Shared knowledge fed by the IoT network is used by smart layers (KB and MAS) to reason and propose an action on one or more IoT devices. The implementation is based on underlying ontologies of comfort and actions, real-world IoT devices and a set of agents representing each actuator. These latter participate to the global system's organization by playing a specific role and by having to reach local goals. Distribution of tasks and emergence of behaviour will result in the selection of actions to execute in order to meet this request for comfort change.