Institute of Information Science, Academia Sinica

Print

AirBox: a participatory PM2.5 sensing system

AirBox: a participatory PM2.5 sensing system

Micro air quality sensing is an emerging sensing paradigm that combines the advances in low-cost air sensors, long-range low-power communication, and big data analysis for environmental monitoring. Comparing to conventional environmental monitoring systems, it is more affordable for large-scale and dense deployment, and it can provide measurements at a finer spatio-temporal granularity. In recent years, a number of large-scale micro air quality sensing systems have been deployed worldwide by different governments, industry, and communities.

We have been devoting to micro air quality sensing research since 2013. In 2015, we launched the AirBox project, which engages citizens to participate in fine particulate matter (a.k.a. PM2.5) sensing and empowers participants to make low-cost PM2.5 sensing devices on their own. The projects grew rapidly with supports from local governments, domestic IT companies, and citizen communities. In 2019, the project also received a 4-year government grant to deploy AirBox devices in every K12 school in Taiwan. By March 2020, we have deployed more than 15,000 devices in 58 countries, and our open data portal has become the largest and most popular repository of micro PM2.5 sensing data in the world.

In addition, we have investigated the inner properties of micro air quality sensing data and conducted a number of algorithms for data analysis. For instance, we designed an anomaly detection framework (ADF) to detect anomalous devices/events in sensing data streams, and we proposed a hybrid model for short-term air quality forecast by incorporating data clustering and neural networks. Moreover, by combining real-time sensing data and short-term forecast results, we proposed a clean air routing (CAR) algorithm to provide the route recommendation of minimal air pollution exposure. Our research results are of both theoretical and practical values and have been published in prestigious journals. The proposed algorithms have been implemented in the AirBox system, and the results have been used by governments and research communities.

The project has also received extensive attention from the media, the public, and international research community. In addition to domestic media coverage, it was covered in “Dust Island – Particulate Matters”, the first documentary focusing on air pollution in Taiwan, in 2019; and it was featured by CBS (USA) and BBC News (UK) in 2018 and 2019 respectively. The project has good international visibility in the community, in terms of tight collaboration with top research teams and government agencies, such as Array of Things project (Argonne National Lab, USA), The COGfx Study (Harvard University, USA), AQ&U project (University of Utah, USA), SmartAQNet project (Karlsruhe Institute of Technology, Germany), DustBoy project (Chiang Mai University, Thailand), and AirBox Korea project (Gyeongsangnam-do Provincial Government, Korea).

Finally, by exploiting the finer-grained data resolution of micro air quality sensing, the AirBox project not only benefits researches in computer science and environmental science, but it also stimulates interdisciplinary innovation in public health, risk management, urban planning, atmospheric science, and science and technology studies. The project has created a positive ecosystem involving academia, industry, governments, and citizens; and it has potential to facilitate smart city, smart environmental governance, and public-private partnership for common good in the future.