HESTIA: A Home Environment Simulator Targeting Inhabitant Activities
Mayki dos Santos Oliveira, Denis Robson Dantas Boaventura, Eduardo Ferreira da Silva, Joel Machado Pires, Isaque Santana Copque, Bruno P. Santos, Ivan Machado, Cassio Vinicius Serafim Prazeres, Maycon Leone Maciel Peixoto, Gustavo B. Figueiredo, Eduardo Santana de Almeida, Hérsio Massanori Iwamoto, Frederico Araújo Dur�ão · 2025
The diverse range of Internet of Things (IoT) devices available today offers great potential for home automation and Smart Home (SH) development, whether partial or complete. However, for research in this area to advance, it is essential to overcome the scarcity of data, as data collection in SHs faces significant challenges, such as the complexity of devices and environments, privacy and security concerns, and restrictions imposed by proprietary devices. Furthermore, freely available datasets often have limitations, such as incomplete annotations, restricted device usage, and insufficient spatial and temporal coverage. To address this limitation, we developed a novel SH simulator that supports multi-state devices and captures detailed information on device usage and inhabitant activities, including metrics like lamp brightness and air conditioning settings (temperature, mode, etc.). The primary goal is to provide researchers in SHs, the SH, and related fields with a tool for generating datasets that effectively represent an inhabitant or group of inhabitant activity patterns in simulated environments. Home Environment Simulator Targeting Inhabitant Activities (HESTIA) resolves the inconsistency in message exchanges between devices, which often stems from differences in device manufacturers, while also eliminating the time-consuming and costly process of collecting real-world data by generating synthetic datasets. The simulator HESTIA offers rich, dynamic information that can be tailored to specific user profiles. To validate our method, we collected data over two weeks from a real environment using four devices and compared it to a synthetic version of the same setup generated by HESTIA. The results show that the simulator accurately replicates real-world environments, even for devices with inconsistent status changes. Additionally, we evaluated the computational complexity of this simulator in terms of time.