22set2025
10:00 Master's Defense room 85 of IC2
Topic on
AirCEP - An application for Air Quality monitoring with Complex Event Processing
Student
Gabriel de Souza Ribeiro
Advisor / Teacher
Luiz Fernando Bittencourt
Brief summary
This work arose from the need to address the challenges of air quality monitoring in Brazil, where infrastructure is limited and operating costs are high. AirCEP was developed to offer an efficient and affordable solution. This solution seeks to reduce network usage, hardware resources such as processor and memory, and latency, while generating alerts for adverse air quality conditions. Air monitoring is particularly necessary in large cities, where air pollution can affect people's health and the environment. In Brazil, there is limited availability of monitoring equipment and few stations capable of measuring pollutants such as particulate matter (PM2,5 and PM10), sulfur dioxide (SO2), ozone (O3), carbon monoxide (CO), and nitrogen dioxide (NO2). AirCEP consists of three key components: a data forwarder, which uses configurable filters to reduce data volume; a data processor, which uses Apache Flink to process data in real time and generate alerts; and a visualization dashboard, created with Grafana, that displays the data and alerts generated. The performance evaluation was conducted in two scenarios: a local environment (components on the same machine) and a remote environment (sensors far from the processor). The results showed that the data filter reduced network traffic by up to 30% and the consumption of resources such as CPU and memory. Latency, however, was more influenced by the physical distance between the sensors and the data processor than by the data filter. The main contributions are: (1) a modular and scalable architecture; (2) reduced network traffic and computational resource consumption; and (3) real-time detection of complex events. AirCEP proved to be a viable solution for regions with low infrastructure availability, reducing network traffic by an average of 30% and CPU consumption by up to 19% compared to traditional approaches.
Examination Board
Headlines:
| Luiz Fernando Bittencourt | IC / UNICAMP |
| Adenilso da Silva Simão | ICMC / USP |
| Juliana Freitag Borin | IC / UNICAMP |
Substitutes:
| Islene Calciolari Garcia | IC / UNICAMP |
| Douglas Dyllon Jeronimo de Macedo | CIN/UFSC |