CUPERTINO (KPIX 5) — Two Bay Area high school students have invented a tool to predict and prevent fires; it may be the firefighting tool of the future.
Aditya and Sanjana Shah, both seniors at Monta Vista High School in Cupertino (and who share the same last name but are not related), have watched with sadness how wildfires have engulfed California. They wanted to use their skills and creativity to create a device that could help.
The students tested their smart wildfire sensor in the Santa Cruz mountains. The device, which was strapped to a redwood tree, uses weather data, cameras and machine learning technology to predict the area in which a fire is most likely to occur.
The sensor collects and transmits weather data, including wind speed, temperature and humidity.
A built-in camera takes real-time images of forest fuel sources and analyzes them with machine learning and artificial intelligence. These tools develop patterns that could spell out potential fire danger.
“Then, each of these models will split up a predictive score of a chance of a wildfire happening, first based on historical weather data and also the fuels. Those predictive scores are combined and it’s able to predict the chances of a wildfire happening,” said Aditya Shah.
“In the last decade we have had a tremendous amount of technology at our fingertips. But we are not effectively using it in order to solve the problems that have been occurring for millions and millions of years,” said Sanjana Shah.
The two students created the device after brainstorming in their environmental science club at school.
“I want to help solve people’s issues by using a product, and I really feel deeply touched by the wildfires. It hit so close to home,” said Aditya Shah.
Theoretically, there would be many sensors pre-installed throughout forest lands, each sending data to the cloud, where it can be downloaded by fire departments as they aim to get ahead of disasters.
“We would like to approach state governments to use this as a nonprofit solution. We’re not planning to sell this to them, we’re planning on using this across the nation,” said Sanjana.
So far, the students say testing has proven the sensor to be 92% accurate. Additional testing is needed, which will happen as soon as this year’s fire season ends.