BRINGING A NEW WAY OF THINKING TO FIREFIGHTING

The map of Australia on fire, pictures of smoke and fire behind cityscapes, images of brave men and women rescuing koalas, videos of helicopters and airplanes battling the flames from above: these pictures dominated our news feeds and our minds during Australia’s bush fire season in 2019/2020. Many sat at home and opened their wallets, desperate to find a way they could help. Jacobo Domingo Gill, COO and Co-founder at Robotto told me that he distinctly remembers the “footage of civilians getting trapped in flames, seeing their houses burned, and having to escape to the beaches.” In addition to this destruction, biodiversity was affected as we saw the Koala habitats get destroyed and its species brought to the brink of extinction.

This same feeling of helplessness and a wish to do more was felt among many of us in the tech community. We looked at the tools firefighters had at their disposal and then at our technology and began a tireless finding useful applications for our technology to assist in the fight.

It’s essential to understand what’s been used in the past and why these tools were used when applying new use cases to existing technology. Assuming shiny new tech is the answer to our problems has been the pitfall of many in our industry. Here I endeavor to take you through some of the thoughts and observations we’ve made on our journey of bringing AI tech to the firefighting world.

From traditional to innovative

Traditional methods for fighting wildfires such as those in the 2019/2020 season are relatively low tech. They involve using water to push back flames from on the ground or in the air. Firefighters also use what is commonly known as burning out or backburning to halt the fire’s progression. The key issue with such methods is that they are reactionary- they work to control something that may have already grown out of control by the time their efforts come to fruition.

Some technology has made its way into the world of firefighting. The use of drones and satellite imagery has become a staple for many departments when battling large wildfires, but the implementation of smart, autonomous technology still lags.

In the 2019/2020 bush fire season, we found many examples of departments utilizing NASA images to determine the size and velocity of the fires. While this is undoubtedly a step in the right direction, the delay in images and the inability to determine the direction of the spread still puts firefighters on defense rather than offense- leaving them once again in a reactionary position, unable to utilize their resources efficiently.

Seeing those images pushed us at Robotto to understand the need for a product like AWRA that could help firefighters stay one step ahead in their fight and help reduce the impact forest fires have on biodiversity.

Prevention and an understanding of ecology

Sometimes it seems as if the only explanation for the increase in the number of wildfires is climate change. While this may be true to an extent, blaming this increase entirely on something we have little control over (beyond reducing carbon emissions, etc.) can be dangerous. Cristina Santin Nuno, Associate Professor of Biosciences at Swansea University, notes that blaming wildfire alone can be dangerous, citing that “it is very dangerous to think it is only climate change making fires worse at that can lead to people thinking ‘there is nothing we can do.’” (Theneweconomy.com) After the fires in California last year, many began to research this very question, is climate change solely to blame for this increase in devastation?

The answer was no. In many ways, our historical treatment of natural wildfires and modern inhabitation of fire-prone areas Is also to blame. On top of this, our way of analyzing firefighting is black and white. We think only in terms of controlling/fighting fires or prevention and leave little contemplation on how we as a society can work to give nature the ability to act while keeping our communities safe.

A new tide – looking towards the future.

As devastating as the fires were in Australia last year, the immense focus and passion that followed this disaster awoke a passion and raison d’être among those in the tech community, including us at Robotto. Institutions from NASA to private companies, including those in the startup community, began looking for ways to immediately impact their technology. Technologies such as satellites, artificial intelligence, and robotics all started finding ways to work to halt wildfires.

This energy is fantastic and is one I know well. Every time I walk into our offices (when COVID isn’t peaking, of course), I’m reminded of this passion and energy. But energy and willpower are, unfortunately, not enough when it comes to actualizing and integrating exciting new technologies. ‘

It’s essential that governments and communities work actively to implement new technologies and begin to think beyond either combating fires or preventing them, moving instead towards combining the two. Jacobo Domingo Gil, COO at Robotto noted that “a path also needs to be opened for technology to break into well-established industries like firefighting. To do this, we need regulation to advance at the same pace as innovation and help us integrate with the already established methods.”

While we perfect our technology, we have to work together with government and private contractors to find a way to implement Robotto’s unique tech to help move firefighting away from being solely reactionary or preventative, giving firefighters a tool that allows them to be in the right place at the right time, and be ahead of the fight.

THE TECH DRIVING ROBOTTO’S BATMAN IN THE SKY

Reconnaissance across industries has begun to rely heavily on drone technology, but the development and innovation of tracking software looks to revolutionize industries from fire safety to powerline inspection. We’ve put together an overview of how our system gets its data. Take a look to learn more about how our unique, specialized tracking drone software equips users with state-of-the-art technology that’s changing the game as we know it.

Efficiency is the name of the game in most industries, and applying AI to drone software makes inspection and reconnaissance easier, safer, and more efficient- but how can we make something more efficient while simultaneously focusing on protecting our environment and staying true to our values? This last question is at the core of everything we do at Robotto. When developing or identifying new use cases, we often take a step back and ask ourselves, is this a responsible application of AI, and does it work towards making industries more sustainable?

It’s questions like this that altered our bachelor’s project’s path, moving our focus from drone software that would deliver our pizza to state-of-the-art software aimed at tracking and equipping firefighters with the information they need precisely when they need it. Pushing back on our inner craving for juicy pizza, we worked to develop software with autonomous features, data analysis, and visualization and a method of providing this data in real-time without delays.

To your average joe, equipping a drone with these features seems like a two-step process, but the development team at Robotto would beg to differ. Our team identified three different algorithms needed to ensure efficiency during the development of this software for wildfire tracking. Using a custom-built database for recognition and processing of fire data, AWRA, the Autonomous Wildfire Recognition and Analytics platform, ensures intelligent processing and visualization of data, providing it to end-users on a topical map in real-time.

Batman, is that you?

From his perch in the sky, Batman views the city, on the look-out for people in need. While he sits there considering his dominion, he notes the buildings' size, the distance between structures and predicts several possible outcomes for the situation developing below. Much like the Batman of our childhood dreams, AWRA views a developing situation from high in the sky, collects relevant data, processes this data, and provides users with proactive data that's ready for use.

Unlike the real Batman, our software relies on GPS location, a monocular camera, and several data collection and analysis systems for fast and accurate calculation of the location, size, and direction of a wildfire. While flying over wildfires, one of AWRA's core algorithms identifies the edge of the fire as well as differences in fire intensity. With this information, the software then calculates the fire's size, giving firefighters the exact size and spread of the fight ahead of them.

Autonomous is the future

While autonomous drones are still operational in many parts of the world due to strict drone regulations, we believe that future drone operations will utilize and rely on autonomous functions, including autonomous flight. From a safe distance, operators select an area on a digital map using two coordinates and two lengths to determine the rectangle's edges wherein the area navigation is to be performed. An area sweep is performed inside the designated rectangle, following a snake pattern, tracking the fire in real-time.

Utilizing autonomous features not only frees up time for the operator but removes aspects of danger from an otherwise risky job such as firefighting. Increasing efficiency and removing said danger allow operators and firefighters to spring into action with increased situational awareness, resulting in faster control of wildfires, helping protect communities and nature.

Just a few more details

We know the process the software takes to provide data in real-time, but we also know that Batman, well he can do a lot of impressive things- the same is true for AWRA. Unlike traditional fire surveillance and detection methods, drones equipped with AWRA software can view areas otherwise unreachable due to no-fly zones, heavy smoke, or night-time flight regulations. In addition to this already impressive improvement from the current situation of fire surveillance and detection, drones operating with AWRA do not rely on an internet connection, giving users the ability to collect and take action on data in remote areas with an accuracy of +/- 0.5 meters.

Testing with the best

Spain has a history of ravaging, determinantal wildfires, which pushed its government to establish a unique, elite squad of firefighters known as the GRAF Bombers. Their know-how and ability to track and analyze fires is renowned throughout the firefighting world. As programmers and AI nerds, we know how to make our technology work, but our knowledge of how fires behave was limited. This limitation encouraged us to reach out to firefighting organizations, including GRAF, letting us test our technology with the best in the field of wildfire fighting.

This is why, amid COVID-19, our team traveled to Barcelona, where we were able to test our software, displaying how drone software such as AWRA increases the situational awareness of firefighters. Used to operating drones manually, the GRAF team was limited in their efforts to extinguish the test-fire as it did not provide them a broader perspective of the direction and size of the fire. Applying AWRA, the team delivered more precise and actionable information about how the fire would spread and evolve, allowing them to spring into action quickly and tackle the fire more effectively.

The development team at Robotto has a lot up their sleeves as they work to innovate and shape the autonomous future. To follow their development, make sure to find us and subscribe on Linked In and Twitter!