Banca de QUALIFICAÇÃO: HEITOR MEDEIROS FLORENCIO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : HEITOR MEDEIROS FLORENCIO
DATA : 21/06/2017
HORA: 09:00
LOCAL: Núcleo de Pesquisa e Inovação em Tecnologia da Informação - NPITI
TÍTULO:

A Probabilistic Graphical Model for Inference of Stability in Wireless Industrial Networks

PALAVRAS-CHAVES:

Wireless Industrial Networks, Network Stability, Probabilistic Graphical Models, Bayesian Network, Random Markov Field.

PÁGINAS: 65
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
SUBÁREA: Eletrônica Industrial, Sistemas e Controles Eletrônicos
ESPECIALIDADE: Automação Eletrônica de Processos Elétricos e Industriais
RESUMO:

During the last years, the use of communication technologies, especially wireless, has been intensifying in the industrial environment. The equipment establishing the control and monitoring systems, such as sensors and actuators, are being developed with wireless communication capabilities in an environment that, due to large levels of interferences, becomes unfavorable for the implantation of networks based in conventional protocols. Thus, since 2010, new protocols have arisen based, in their majority, in the IEEE 802.15.4 standard, for the incorporation of the wireless technology in industry level. Those protocols, such as IEC 62591 (WirelessHART), IEC 62734 (ISA 100.11a) and IEC 62601 (WIA-PA), have been conquering, slowly, confidence from the market, as they guarantee high network stability, low latency transmissions and low energy consumption. In this context, it is essential to develop tools for analysis and diagnostics that assist the manufacturers, designers and network users.

 

Nevertheless, in order to analyze the performance of those systems, the use of modelling techniques which are, generally, based in empiric models, is necessary. Incomplete or imprecise collected data acquired to build an empiric model can result in unsatisfying results, affecting the performance of its own analysis. Those situations, that is, in the face of uncertainty, probabilistic methods are more suitable for classification and prediction. Probabilistic techniques are able to incorporate the relations between the system variables that, perhaps, would not be evident in the data sample in non-probabilistic models.

 

Hence, this thesis proposal presents an inference model of stability in industrial wireless networks based in several collected network parameters, such as power transmissions and reception levels, quantity of communication links, transmissions rates and redundancy.  Given the non-deterministic characteristics of industrial communication systems, the inference is modelled with the support of two probabilistic graphic models techniques: Bayesian Network and Random Markov Field. The Bayesian network performs a local inference of the stability in the communication paths and the Random Markov Field performs a spatial analysis to infer the network stability. That stability inference contributes to the development of packet routing techniques, instrument allocation, energy consumption reduction and, consequently, enhancement of network reliability and availability.

MEMBROS DA BANCA:
Presidente - 347628 - ADRIAO DUARTE DORIA NETO
Interno - 1837240 - MARCELO AUGUSTO COSTA FERNANDES
Externo ao Programa - 350241 - JORGE DANTAS DE MELO
Externo à Instituição - DENNIS BRANDAO - USP