Time to send robots and drones to combat wildfires

Slide Menus

Show/Hide Left Slide Menu Show/Hide Right Slide Menu Show/Hide Top Slide Menu Show/Hide Bottom Slide Menu City Times
WKND
Inspired Living
KT Home ICC World Cup 2019
Photos
Videos
Interactives
WKND
Young Times
Inspired Living
Dxbuzz
Cyprus
Flowers of Tolerance
UAE National Day
India Reality
Kartarpur Corridor
Sign in using OR Forgot Password? Register Now! Terms | Privacy

There is an unequivocal need for humans to be entirely replaced by machines in hazardous scenarios.

We saw some horrific pictures of the monstrous fires that swallowed up large swathes of forestland in Australia. There has been a constant stream of poignant photos of rescued humans and animals, and heartbreaking stories of those who could not make it. Australia was literally waging a war against what grew into an inferno of epic proportions.

Last year, Amazon was burning in a similar fashion and armies were called in to curb it. In the US, California has been a hot bed for wildfires every year. In US alone, for the past two decades, an average of 72,400 wildfires scorched roughly seven million acres of land every year. Such wildfires have doubled since the 1990s.

Tech Pundits are trying to find ways to preempt such occurrences – the same way that today UAE's met department is able to issue a bad weather warning so that children can skip school. We take such predictions for granted today. But getting weather forecast right has been a huge human feat on the back of decades of weather big data and centuries of human endeavour. Even in 650 BC, Babylonians were doing weather forecast based on cloud patterns. When I was young, we had little confidence in the daily weather forecast. Not anymore. Fire behaviour analysts are looking for similar mathematical models that are equally accurate.

In early 1970s, Dick Rothermel created the first mathematical model to predict the spread of fire in the US. He burnt fuel in his wind tunnel and controlled factors such as wind speed. He plotted his observations on his graph to eventually work out a set of equations that could be used to predict the spread of wildfires. The Rothermel model became the basis for fire behaviour analysts to better predict how a blaze might spread. They are in effect trying to reverse engineer a forest fire.

Sounds simple? Not quite.

There are so many factors that affect a wildfire in the US alone, let alone other countries. For example, the chemistry of plants gives us insights into their propensity to, quite literally, add fuel to fire. Remote sensing data from satellites has limitations because it cannot differentiate between smoke and cloud. So, scientists are hard at work to narrow the gap in their understanding.

They are layering the Rothaermel model with many more real-world factors to further adjust the model. Simply put, they are trying to understand the path of the fire and the time it would takes to spread in those directions.

Some institutions model the movement of fluids and gases in a wildfire. This computational fluid dynamics (CFD) model divides an area into a chessboard or a grid, and assesses how each grid would interact with others. This model requires massive computational power, which means it could take days to predict. No wildfire would be willing to wait that long.

Therefore, fire behaviour analysts are trying to find a simpler model that could work on the laptops they carry into remote forest areas. They also need to win the trust of firefighters.
There is always a tussle between human experience and mathematical models. Yet, with so many new factors playing out and the scale of wildfires being unprecedented, experience is increasingly relying on such predictive models.

But we need more than that to fight blazes that spread thick and fast. Drones can come in handy to take aerial pictures, establish connections with firefighters, and douse fires.

Drones and robots can easily operate in smoke-filled areas. Smokebot in Germany, for instance, is able to gather data where firefighters cannot. Mexico-based Descartes Labs has a tool that uses satellite images derived from both visible and infrared light to trace the heat signatures of fires in the night and on cloudy days. It shares the information as time-lapsed videos with hashtags.

We need a perfect storm, a convergence of many types of technologies to prevent and arrest wildfires. A more collaborative approach by the industry can spark new approaches.

For example, many conservation technologies already used for protecting forests can be used. Drones that perform aerial reconnaissance to detect illegal mining in forest areas can double up for gathering data for forest fire prediction. Conservation agencies are using high-tech acoustic monitoring devices to detect chainsaw sound and gunshots. These can have sensors for capturing the crackling sound of burning foliage. Crowdsourced Instagram and Facebook photos are being layered on top of maps. Wildfire predictive models have a lot to learn from how weather forecast became more accurate, in addition to the obvious correlation between weather conditions and forest fires.

We need more of intelligent drones, robots, and sensors to prevent wildfires. The courage and hard work of our firefighters cannot be overemphasised. But the scale of the Australian and Amazonian wildfires has exposed obvious limitations and risks of humans involved in firefighting. There is an unequivocal need for humans to be entirely replaced by machines in such hazardous scenarios. This is possibly the one industry in which humans need to be fired from their frontline jobs and replaced by intelligent machines.
Shalini Verma is CEO of PIVOT technologies

ERROR: Macro /ads/dfp-ad-article-new is missing! MORE FROM Opinion and Editorial