INF - Instituto de Informática
URI Permanente desta comunidade
Navegar
Navegando INF - Instituto de Informática por Autor "Abreu, Marcos Felipe Barboza de"
Agora exibindo 1 - 1 de 1
Resultados por página
Opções de Ordenação
Item Detecção online de dispositivos sem fio intrusos usando o sinal eletromagnético de transmissão(Universidade Federal de Goiás, 2022-09-23) Abreu, Marcos Felipe Barboza de; Vieira, Flávio Henrique Teles; http://lattes.cnpq.br/0920629723928382; Cardoso, Kleber Vieira; http://lattes.cnpq.br/0268732896111424; Klautau Júnior, Aldebaro Barreto da Rocha; Corrêa, Sand Luz; Cardoso, Kleber Vieira; Vieira, Flávio Henrique TelesThe identification of Internet of Things (IoT) devices through the electromagnetic signal is a topic widely investigated in the literature, and this technique is considered highly accurate by several works. The use of offline techniques, that is, when there is no presence of new devices, is widely explored, but so far, systems are not found effectively using the detection of unknown devices in the online way, i.e. , one of the greatest potentials of this type of technique has not been investigated. This work presents an online system that differentiates authentic devices from intrusive devices. For this, the use of the probability matrix of classifiers is explored, aiming to identify unknown devices by them. In addition to the technique, it is also presented a system that features a modular, extensible and generic architecture, which aims to minimally interfere with the normal flow of an Internet of Things application. The system is implemented using the GNU Radio tool and experiments are presented, which aim to show the feasibility of the technique. The entire discussion is based on data collected from real environments, using devices from wireless communication technologies LoRa and ZigBee. In addition, the work analyzed data from WiFi technology, from collections found in the literature. Tests show that it is possible to identify unknown devices in the order of milliseconds, with a low error rate.