How Artificial Intelligence Can Tackle Indoor Air Pollution

The COVID-19 pandemic has forced countries across the globe to enter into lockdowns in much of 2020. While it introduced many economic challenges, one upside was that the world saw a reduction in outdoor air pollution. With restrictions on global and local travel, shutdown of schools and businesses, and halting of industrial activities, there has a been a substantial reduction in nitrogen dioxide (NO2), modest reductions in fine-particle air pollution (PM2.5), and higher levels of ozone (O3) according to the State of Global Air 2020 report. In addition, global emissions from burning fossil fuels fell by a record 7% in the same year according to the Global Carbon Project, an international group of researchers who track emissions.

However, as countries have begun to open back up, pollution levels have rebounded to pre-pandemic levels. Scientists at NASA noticed NO2 bouncing back to near normal levels in China three months after the country shut down transportation and much of its economy in early February 2020. Such a focus on outdoor air pollution has diverted attention away from indoor air pollution, to which exposure is much greater. According to the Environment Protection Agency (EPA), indoor air pollutants levels can be 2 to 5 times higher than outdoor levels, with lockdowns across the globe further exacerbating this. With society already spending close to 90% of time indoors prior to the shift towards working from home and generally more time spent at home, the issue of indoor air quality becomes ever so important. With that, solutions that make use of technologies such as Artificial Intelligence (AI) have come into the picture with the aim of alleviating the problem of indoor air pollution. These innovative solutions include Creo’s AI-powered indoor planting system bringing more nature indoors and aeris’s AI smart domestic air purifiers among others, all looking to improve the poor indoor air quality faced by many on a daily basis.

One way in which indoor air pollution levels can be lowered is through the addition of an indoor living wall or vertical garden. Plants significantly improve air quality because they act as natural air purifiers. Creo, a green technology company based in San Francisco, recently unveiled a smart, green living system, in the form of a technology platform, called the BioBulb. The system provides a sustainable solution for growing nature indoors and in environments where resources are scarce. It uses up to 80% less water and 50% less energy to operate through its precision feeding function whereby water and nutrients are only provided when needed. BioBulb makes use of Creo’s AI software, called Darwin, which is coupled with three different groups of sensors, to learn the optimal growth patterns of each plant in various scenarios and optimize the number of resources allocated, while also providing insights into future growth of the indoor plants. The BioBulbs are connected to a Bio-Server, which houses all the software and technology and is attached to the building’s water supply. The sensors measure temperature, pH levels, humidity, and nutrients, and monitor light density, the leaf surfaces, and the surrounding environment. Additionally, a computer vision camera is equipped to detect the plant’s growth. Using data from sensors, BioBulb encourages the growth of plants in environments with suboptimal conditions in a precise manner. Creo’s system can be integrated with existing smart home systems and appliances such as Google Nest and Samsung that provide additional data points of the home and further improve on the optimization process.

Aside from natural air purifiers, companies like aeris, a Swiss startup that creates commercial and residential air purification systems, develop AI-powered smart air purifiers. These purifiers learn and adapt to each individual’s personal habits and the air quality of the home using data collected from smart sensors. With the use of integrated ML algorithms, the purifier learns over time to be more efficient, reduce noise and power consumption, and ultimately extend the life of the filter. With the development done by MIT engineers, the company’s medical-grade purifiers have the ability to remove at least 99.95% of indoor pollutants. Driven by AI technology, these air purifiers can provide users with enhanced indoor air quality.

Nonetheless, ABI Research believes more can be done with regards to data processing, power consumption, and privacy. According to the ABI Research market data Artificial Intelligence and Machine Learning, it is forecasted that the total shipments for TinyML devices in smart home devices with edge AI chipsets will increase to 513 million in 2025, from 195 million in 2020, growing at a 5-year CAGR of 14%. This provides indication of the growing trend of ultra-low power AI integration into smart home devices such as air purifiers and indoor plant systems mentioned above. Sensors equipped with TinyML chipsets are able to provide feedback to end users based on the changes to their surrounding environment, including drastic fluctuation in humidity and temperature, absence of light source and specific nutrients, and sudden spike in harmful air particles, without major impact on the overall power consumption. Most important, these raw data from personal spaces are processed locally by the sensors, so only metadata or processed data are being sent to the local smart home systems or the cloud for further analysis.

Indoor air pollution is a leading risk factor for premature death. According to the Institute for Health Metrics and Evaluation (IHME), it is estimated that indoor air pollution is responsible for 1.6 million deaths each year. Such a serious issue requires more attention and investment to allow for further innovation in solutions to help improve lives through improving indoor air quality. Creo’s BioBulb and aeris’s AI-powered air purifiers have provided a taste of how AI can be used to tackle the issue of indoor air pollution. This can be expanded to other forms of pollution such as outdoor pollution and noise pollution with the proliferation of sensors to measure all kinds of data.