Houssam KANSO

Research interests

• Industry 4.0

• Cyber-Physical Systems

• Internet of Things

• Green IT

Projects

Thesis

Title

Contributing to the Energy Efficiency of Smart Homes: An Automated Management Framework

Description

In recent years, the power consumption of Cyber-Physical Systems (CPS) has been increasing due to the increasing number of connected devices (e.g., smart appliances, plug-and-play IoT devices...), mainly in the residential sector. A large number of devices integrate sensors allowing them to produce data describing the state of a device or the behavior of a person (e.g., temperature sensor, presence sensor...). In addition, numerous devices are equipped with actuators capable of accomplishing tasks impacting the environment (e.g., light control, heating, ventilation, air-conditioning system...). These devices have the potential of collecting a large amount of data that can be useful for power estimation and management. However, current energy management approaches are mostly applied to limited types of devices in specific domains and are difficult to implement in other scenarios. They fail when it comes to their level of autonomy, flexibility, genericity, monitored metrics, and heterogeneity of studied devices. To address these shortcomings, we present, in this thesis, an energy management approach for connected environments based on generating power estimation models, representing a formal description of energy-related knowledge, and using reinforcement learning (RL) techniques to accomplish energy-efficient actions. We illustrate our proposal in the smart home domain. We first present an automated power modeling approach used to generate accurate real-time power estimation models for any type of devices in heterogeneous environments. Then, we present an energy-oriented extension for a reference ontology. The latter aims to represent useful concepts used for energy management purposes in connected environments. Furthermore, we develop algorithms of our RL approach that exploit knowledge from both the power estimator and the ontology, to generate the corresponding RL agent and environment. We also present different reward functions based on user preferences and power consumption. The proposed approach performs well given the low convergence period, the high level of user preferences satisfaction, and the significant decrease in energy consumption. The main contribution of this thesis is to guarantee autonomic management of energy consumption. It also provides visibility on energy drains by estimating the power consumption of devices in an automated manner. And lays out a way to represent energy-related knowledge. Finally, it ensures that energy-efficient actions are executed in heterogeneous environments.

Members

Publications