Ambee :Mapping India’s air quality using sensors & AI

Spread the love

After discovering that the particulate matter in the air around him was 32 times the prescribed levels, Ambee  Madhusudan Anand realized it was time to combat the daunting challenge of poor air quality, and turned to data science & APIs to accomplish this.

There’s a concept called Unintended Consequences that’s broadly explored in social sciences, and it simply means unexpected or unforeseen outcomes of purposeful action. While the difficulties brought on by the COVID-19 pandemic cannot be undermined, it threw up many unintended consequences. An unexpected consequence of the summer lockdown was that air quality began improving in several cities, like in Jalandhar, Punjab. 

16 out 18 most polluted cities in the world are in India. 78% of the population are at high risk for pollution-related ailments. Traffic, garbage burning, crop burning, construction activities, factories collectively choke our cities every day. The WHO states that air pollution is among the leading environmental health risks today. And Madhusudan Anand experienced this first-hand. After moving to Bengaluru with his family, his six-month old son would often wake up at night, struggling to breathe normally. Despite being administered medicines, the child’s symptoms would flare up in a few days. He wondered what was in the air that was making his son so ill.

According to the data collected at the nearest air quality monitoring station (which itself was 13 km away), the air quality was around 25 ug/m3 – supposedly deemed normal. But Anand was not convinced. He built a rudimentary sensor that began tracking air quality in his neighbourhood. The sensor showed that particulate matter near his residence was 800 ug/m3! The reason for these shockingly high levels? A garment factory, combined with the general pollution of the area, which houses some of Bengaluru’s biggest companies and factories.

Along with Akshay Joshi and Jaideep Singh Bachcher, Madhusudan Anand started Ambee in 2017 – aimed at measuring air quality at a hyperlocal level by collecting multiple kinds of data, and providing this information in real time. Ambee is the only company in the world that’s collecting air quality data in such a granular manner.

“The WHO recommends that an Air Quality sensor should be installed every square kilometer. But, India has only around 220 such monitors. In addition to our proprietary sensor network, we heavily rely on data from satellites and government organizations to provide real-time, hyperlocal air quality updates,” explains Anand.

Their proprietary technology AIONN-MetNet (Aggregation & interpolation of Meteorology data over Neural Network), which combines a model trained on decades of historical weather data, weather radar data, satellite information and Gaussian interpolation that provides upto a 200m² resolution of weather data for the whole world geospatially. An inhouse training model utilises this data to accurately predict the quality of air. These sensors are EPA benchmarked in addition to having local certifications like NABL and CEE. Ambee has installed 100 such sensors in Bangalore, and are on their way to install 500 across India.

Another unique feature that Ambee offers are its APIs. This is a standard subscription model that various businesses can utilise depending on their need and include weather, air quality, water vapour, fire, soil and even pollen. “This makes the acquisition of hyperlocal air quality data more streamlined and accurate,” says Anand.

Ambee has collected air quality data of over 100,000 pincodes; and is essentially creating a GPS for India using air quality data, adds Anand. This real time data is available on their website called writing this story, I just checked the air quality in my neighborhood in Bengaluru and it stands at 31, marked as ‘good’.)

Varna Bhat : Rahasya Vodka Founder

Ambee is among the handful of companies that can be called homegrown. Anand says he and his cofounders put together the first sensor using a LinkitOne Grove starter Kit and MediaTek on an Arduino board. Ambee had an AQI sensor, CO sensor, temperature & humidity, dust, noise and UV sensor hooked to a Bluetooth and a Wifi module to send data to the web server that would show all of this data on a web application on a dashboard. This was built at an IoT hackathon at the Amazon Development Center in Bangalore, which they eventually won. But when they decided to pursue this product more seriously, Anand and his friends faced a fair amount of trouble sourcing hardware. “There are only 35-40 PCB manufacturers in India, which is not enough to support the growth of hardware startups. Despite the constraints, we were keen on developing our sensors in India – we wanted this to be a Make in India product. Homegrown sensors can be built to record local data. Unless we mine local data, we cannot solve problems at scale for India,” adds Anand. After several iterations and improvements, with support from partners like Bosch, Ambee’s sensors are cost effective (priced at Rs. 1 lakh per sensor), solar-powered, permits firmware upgrades, processes data on the edge and are durable.

The Bengaluru-based startup has piqued the interest of marquee investors like Rajan Anandan of Sequoia Capital, Techstars, Touchstone Ventures LLP and actor Aishwarya Rai. The trio believes COVID19 has reignited the debate on climate change, and the company has its sights set on the UN SDG for combating pollution at a global level using technology.

Leave a Reply

Your email address will not be published. Required fields are marked *