Detecção de Anomalias em Cenários de Cidades Inteligentes Baseados em Internet das Coisas

• Ivan Zyrianoff
• Fabrício O. de França
• Carlos Kamienski

Resumo

The deployment of Internet of Things (IoT) based applications for Smart Cities emphasize the need for solutions that can automatically detect events of interest in these systems. Some scenarios, such as smart public lighting and user monitoring in the so-called digital squares, are examples that generate a massive amount of data. In these scenarios, anomalies need to be identified as fast as possible, since corresponding actions should be executed in a timely manner. An atypical event can have huge impacts on the nearby population. A fog computing based architecture has been proposed to make real-time anomaly detection feasible. The model was evaluated with real data from digital squares from the city of São Paulo and with data from synthetically generated sensors, obtaining satisfactory results in both cases.