‘WE CAN SAVE A LOT OF LIVES’
Danbury brothers invent toxic gas detector using artificial intelligence technology
Adam and Ibrahim Bhavnagarwala lit a cigarette in their family’s garage. It wasn’t for the Danbury High School seniors to smoke. It was to test their invention, which uses an artificial intelligence algorithm to better detect harmful gases.
No one in the family smokes. The twins inserted the cigarette into a compressed ball syringe, lit it and then gathered data.
“That was the hardest part of this entire project,” Ibrahim Bhavnagarwala said.
The twins said their device is a success and can sense toxins, such as vape or cigarette smoke, faster than conventional smoke and carbon monoxide detectors.
They created a company, ExaSense Technologies Corp., and are awaiting approval of a patent for the invention.
“These AI-driven algorithms can identify unique patterns in actually any chemical, not only just cigarettes,” Adam Bhavnagarwala said.
This technology could be used one day in hospitals, hotels, pharmaceutical labs, wastewater treatment facilities and other places to detect toxins and help people vulnerable to second-hand smoke, the twins said.
“If we can administer this to hospitals or hotels, we know we can make a significant difference,” Adam Bhavnagarwala said. “We can save a lot of lives.”
They presented their research virtually last month at an MIT conference largely for undergraduate
students and are scheduled to speak at SensorComm 2020, a virtual international conference geared to industry professionals.
“I’m so proud of them,” said Eman Beshtawii, their mom.
The twins developed an interest in computer science in middle school and started taking online classes in artificial intelligence a few years ago. They see the field as growing rapidly.
“It’s taking over our lives, whether that be self-driving cars or, as we’re doing, detecting toxic gases,” Ibrahim Bhavnagarwala said.
Over the summer, they started working on an algorithm that could sense toxic gases for Danbury, he said.
“We wanted to just provide the city of Danbury with a real-time data on toxic fluid concentrations around the city,” he said.
Eventually, they pivoted to cigarette and vape smoke. For about three weeks, they spent roughly five to six hours a day in their garage testing the technology.
Within 300 seconds, their prototype can detect a single lit cigarette in a room as large 7,000 cubic square feet. It’s also better at determining the source of smoke, unlike conventional detectors, they say.
“It (conventional detectors) cannot distinguish whether the source is from burning food or another toxic gas,” Adam Bhavnagarwala said.
Danbury has been a “pioneer” in this industry, Ibrahim Bhavnagarwala noted. The company National Semiconductor was founded in Danbury in 1959, two years after Fairchild Semiconductor Corporation was created in California, spawning what is now known as Silicon Valley.
The twins, who have started applying to colleges and plan to pursue computer science and applied mathematics, hope to encourage their peers to study this field.
During their junior year, the twins started the AI Club and the IoT Club at the high school. They are also urging the district to create a computer science pathway for the proposed Danbury Career Academy.
“The requisites to actually learn these extremely rapidly growing industries, such as AI and computer science, is actually easy,” Adam Bhavnagarwala said. “There’s not much prerequisites, other than basic adding and subtracting.”
Beshtawii said her sons are not “Einsteins.” With passion for their project, any kids could do the same.
“By the time they’re in high school, maybe they’ll even have greater inventions than what Adam and Ibrahim are doing,” she said.