Last year, the USA experienced it’s worse fire season in more than a century, and the epicenter was California. This is somewhat ironic, because one of the primary industries of that state – the technology sector that is concentrated in Silicon Valley – is pioneering new ways to prevent fires, fight them, and save lives.
Firefighters are no strangers to technology, and firefighting robots have been around for a decade. These new approaches, however, focus on the link between AI and robotics. The central idea is that, if AI is already being used to protect governments and businesses from cyberattacks and other forms of seemingly unpredictable risk, then these same AI systems can be used in the fight against fires.
In this article, we’ll look at how AI and robotics are being used to fight fires around the world and investigate what the future holds for these technologies.
The next generation of fire-fighting robot
First up, it’s worth noting that the idea of fire-fighting robots is not, in itself, a new one. Robots have been used to fight fires for more than a decade, and have saved many hundreds (perhaps thousands) of human lives in the process. What is different about the robots being developed and released today is that they do not rely on human controllers: instead, they are able to drive themselves.
The fact that this is such a huge advance points to an often misunderstood point about fire-fighting robots. Most people, when asked to think about the design limitations of such robots, will highlight the obvious danger they face – fire, and the fact that it is extremely hot. In reality, however, we’ve been able to build fire-resistant robots since at least the 1960s (thanks, in large part, to the space program). For most robots, the problem is that their human operators can’t see through smoke.
Figure 1. Time is of the essence for firefighting teams. AI and robotics tech is essential for making rapid life-saving decisions (Source: Freepik)
The difficulties with building fire-fighting robots has not been one of hardware, but of software. Now, with approaches such as continuous integration and continuous delivery able to roll-out new software to these robots as it is developed, these robots might become truly able to think for themselves.
Towards true intelligence
The advent of truly intelligent fire-fighting robots is, as of 2020, some way off. However, there are rapid advances being made in the deployment of AI on edge devices, and a significant level of investment is pouring into the industry: It’s been estimated that AI will add $15.7 trillion to the global economy by 2030. This, in turn, has led to a great deal of optimism about the eventual impact of AI. In a 2020 Digital Skills Survey, 78% of professionals agreed that AI is the technology that will have the most impact on data science for the upcoming years.
This optimism is not based solely on hype. There are real systems, deployed today, that make limited use of AI to help first responders navigate hazardous environments. As Karen Panetta, an IEEE fellow and dean of Graduate Education at Tufts University’s School of Engineering, told Design News recently, “our research utilizes artificial intelligence to help in these dynamic situations by providing real-time image enhancement capabilities to remove smoke, fog, rain, snow, and to leverage recognition algorithms to bring attention to objects of interest, such as humans, animals, and other physical objects. This will help responders better navigate hazardous environments.”
Figure 2. AI uses data to alert responders and nearby civilians of nearby emergencies and keeps this sensitive data protected (Source: iStock)
Similar research projects are taking place throughout the country. AUDREY, for instance, is a software application being developed by National Aeronautics and Space Administration (NASA) Jet Propulsion Laboratory (JPL) with funds from the Department of Homeland Security (DHS) Science and Technology Directorate (S&T) that performs data fusion and provides tailored situational awareness to first responders. The idea behind the system is that it can take historical data on fires – both from real wildfires and test burns – and use an AI to better predict how future fires will develop and change.
Prevention and response
AUDREY, in fact, forms the first of a number of projects that are aiming to use AI to improve disaster resilience – not just preventing and fighting fires, but also managing healthcare systems and coordinating response teams. Bringing us back to where we started, the state of California is collaborating with Microsoft in one of the largest of these projects, which represents the most advanced deployment yet of AI and robotics in order to prevent and respond to disasters.
Take a look through the documents on this new program, and three elements of it will be striking. One is that the “robots” being used here are not the anthropomorphic “androids” one might imagine. Instead, the program makes use of a huge network of IoT sensors, both on the ground and using remote sensing, to feed data to its AI systems.
The second exciting development of this program is that it is taking a genuinely collaborative approach to the deployment of AI. Microsoft have recognized, for instance, that one of the primary values of AI systems is that they can reduce human error. Consequently, instead of trying to replace human operators with AIs, in this approach AIs are used to provide informed predictions to human managers, who have the final decision on what to do about them.
Finally, the program explicitly recognizes that many citizens (particularly in California) will have legitimate concerns about how their data is being collected, used, processed, and protected. Alongside the AIs being used for disaster response, therefore, the program also explores how AI can improve cybersecurity, and therefore protect the same data that it is using.
Ultimately, the goal of these developments is not, necessarily, to remove humans from the process of fighting fires. No matter how “intelligent” an AI is, disaster response includes moral and ethical decisions that only a human can make. However, it’s becoming increasingly apparent that AIs and robotics, due to their ability to turn big data into smart data, can help humans to make these decisions by providing them with the kind of accurate, real-time intelligence they need.
|Ludovic Rembert is a security analyst, researcher, and the founder of PrivacyCanada.net. He spent his career as a network security engineer before taking up freelance writing on a variety of technical and cybersecurity subjects.
- A digital nose gas sensor with AI
- Low-power devices could listen with a silicon cochlea
- The missing tech foundations for smart buildings
- What’s driving AI to the edge
- The role of artificial intelligence in autonomous vehicles
For more Embedded, subscribe to Embedded’s weekly email newsletter.