Middle-schooler Ryan Honary tapped TCS’ Ignite my Future program to create the early wildfire sensing module.
The module includes numerous fire detectors in high-risk areas and connectivity among firefighters through existing cellular network.
Honary enhanced the model with a real-time fire growth predictor using machine learning (ML) algorithms.
Innovate to save forests
Wildfires in the US have been destroying millions of acres of forest land every year, and the incidents have not only been increasing air pollution but have also been extensively damaging ecosystems in several regions. The depletion of forest area has crossed 10 million acres annually about thrice since 2006, with California generally experiencing the biggest chunk, followed by the states of Texas, Colorado, Arizona, and Idaho.
This alarming loss to the environment prompted Californian teenager Ryan Honary to develop an Internet of Things (IoT)-based early wildfire detection system which could save vegetation and the lives of people and animals.
Robust network of sensors pivotal
The idea struck him as he witnessed devastating wildfires in November 2018 that rendered thousands homeless. He also learned that existing detection systems, such as high-end cameras on geological stations or satellites only, were expensive, difficult to be deployed widely, and did not provide early detection.
The following year, Honary grabbed an opportunity at TCS’ Ignite Innovation Student Challenge to build a low-cost wireless field module—“SensoRy AI”—to detect wildfires in their initial stages and minimize the damages they brought on.
The detection part of the module is connected to existing cellular networks which would help notify the authorities concerned in order to prevent the spread of wildfires such as the ones that had devastated California the summer before.
Honary’s innovative solution for early wildfire detection and its containment relied on two vital requirements. “I realized there were two parts to early detection—the first was to place many fire detectors in high-risk areas and the second was to allow connectivity with firefighters,” he said. High-risk areas refer to spots with high voltage equipment.
Ryan’s system also boasts ease of deployment of a robust network of wireless sensors and only a few nodes needing cellular coverage. He got the idea from the internet in its initial stages where all the computers would talk to each other, and even if one was down, the rest would continue to communicate.
When the wireless fire detectors sense flames, the message is relayed to firefighters via an app on their phones, allowing them to respond immediately.
ML-based enhancements and more
The seeds of innovation that were sown in 2019 that helped him bag the Grand Winner position in the Ignite My Future in School (IMFIS) program, a TCS Empowers initiative, germinated into an improved project that includes additional features to help manage such hazardous situations better.
IMFIS is a one-of-a-kind initiative that aims at transforming the way students across America learn through computational thinking, Ryan’s ultimate goal was to build an ML-based system that can predict not only the start of fire but also the growth pattern and speed of the wildfire and provide a potential fire map.
The module planned by Ryan had five detectors: one each for temperature, humidity, carbon di-oxide, wind, and infrared. The sensors were calibrated to trigger only in the presence of an actual flame, as opposed to any type of heat including sunlight.
The fire detectors communicate with each other and with mini meteorological stations. In case of fire(s), the sensors transmit the message to the mini meteorological stations, which in turn transmit the information to the phone app, notifying the user of the situation.
The meteorological stations, using ML, also draw potential growth maps of the fire in real time with the help of all the sensor data captured up until that very point in time. The same is displayed on the mobile app, providing firefighters with an approximate idea about the extent of the potential environmental damage.
Using the technology to address other threats
This hazard management solution, which includes the use of sensors and real-time message transmission capabilities, can also be extrapolated to other disasters like COVID-19. For instance, Ryan designed a fever-detector by replacing the fire sensors with two cameras, regular and infrared ones. The cameras are pointed at a person’s face, and a screen displays the outputs from cameras.
The module also has facial recognition, detects the person’s forehead and temperature, and displays the same on the screen.