The Modeling of a Multi-Agent System for a Domotics Platform
| Authors: | Ferreira Rosa Paulo Fernando, Instituto Militar de Engenharia, Brazil
Tanaka Botelho Wagner, Instituto Militar de Engenharia, Brazil
Santos de Lima Sandro, Instituto Militar de Engenharia, Brazil
Carrilho Antonio, Instituto Militar de Engenharia, Brazil |
|---|
| Topic: | 3.2 Cognition and Control ( AI, Fuzzy, Neuro, Evolut.Comp.) |
|---|
| Session: | Multi-Agent Systems |
|---|
| Keywords: | smart house, pervasive computing, footstep sensor, multi-agent systems, neural network |
|---|
Abstract
The smart house is an environment where several systems will interact autonomously, performing tasks, that in regular houses are done by their inhabitants. The house proposed in this paper is capable to: identify its inhabitants and adapt the rooms according to their preferences; maintaining optimal levels of comfort, safety and energy saving. To act in a personalized way, the house must tell its inhabitants apart, so we propose an identification system that uses a pervasive footstep sensor and a neural network, which are able to identify the inhabitants even when they present variations on their gait patterns. In this project, it’s used the a simple agent communication infrastructure that facilitates the implementation of the distributed agents in an agents’ society, combines MaSE and UML (unified modelling language).
