Sistema de identificação automática de riscos hidrometeorológicos com retroalimentação e reestruturação autônoma da infraestrutura de comunicação

Abstract

The Internet has become essential for the population in different aspects of life, facilitating the use of various services in people’s daily lives. For this reason, it is critical to have a reliable and resilient communication infrastructure. However, one of the factors that contribute to the weakening of this infrastructure is natural disasters, such as floods, which can damage the physical means of transmission and/or make its operation unfeasible and can lead to the unavailability of important services. Using modeling based on outliers’ detection, this course conclusion work aims to identify catastrophic events that can cause failures in communication links and, based on this identification, offer an autonomous infrastructure restructuring strategy. The proposal defines feedback between event detection and network restructuring, in order to allow identified infrastructure failures to help identify catastrophic events. The model input is meteorological data, flooding from Curitiba, in addition to simulated data on the availability of Internet access. Based on this model, it is expected to make the network infrastructure that connects the healthcare units in Curitiba more robust.