Generative AI + Autonomy: The Next Leap for Industrial Drone Programs

Generative AI-run systems point to the future of drone applications

By DRONELIFE Features Editor Jim Magill

Recent advances in the development of generative artificial intelligence (AI) and pending new regulations regarding UAV operations are likely to pave the way for widespread industrial adoption of fully autonomous drones, according to a speaker at the Energy Drone/Robotics and Industrial AI Forum held recently in Houston.

“AI, especially generative AI, is very different from the machine learning that we have been used to,” Nitin Gupta, founder and CEO of San Jose, California-based Flytbase, said. “This is completely transforming the way we look at our data or understand data, and generate insights from that data.”

New applications of generative AI — which can create new content, such as text, images or videos, based on patterns learned from existing data — “is going to create a massive disruption from whatever we used to do a year ago,” Gupta said.

“And all this has come together very recently. It’s less than a year old right now,” he said. Combining the use of generative AI with a system of docked, autonomously operated drones will allow industrial companies to conduct routine tasks such as inspection and maintenance proactively, by letting the system learn and figure out what needs to be done.

Gupta said Flytbase, which develops autonomous drone systems for diverse commercial applications, has worked across a wide spectrum of industries – such as oil and gas companies, utilities, railways and solar companies.

“We understand how to put that scaffolding in place and how to make generative AI really get you the value that you need in terms of inspection,” he said. “You just say that you want these inspections to be done, and the system tells you, ‘Okay, these reports will be ready in the next half an hour or these will be ready in the next two hours,’ and that’s it.”

The system will care of all the scheduling, mission management and flight operations, even down to deciding which docked drone to deploy, based on which drones are equipped with the right software to best perform the job. And, because it is fully autonomous, the operation of the system is no longer the exclusive domain of a company’s drone pilots, Gupta said.

“This is used by your inspection team or by your operations team,” he said. “Because now they are able to just say, ‘What is the job that I want to get done?’ And internally, everything is managed and scheduled by the system, and you get the results and outcomes that you want.”

In the meantime, drone pilots and operators are still in charge of ensuring that the system is running smoothly and are able to take full control in case of an emergency to make sure that everything is operating safely.

Initiating a drone program

Companies initiating their own drone inspection or maintenance program from scratch have to make a number of decisions as to what type of drones and software to use and how to integrate the drone system with the rest of the business enterprise.

“And then there is: how to hire pilots, how to get your certifications and permissions to operate, how do you train your pilots? There are just many moving pieces, which all have to come together before you can build your own program,” Gupta said. “This really slows everyone down.”

It takes most organizations anywhere from six to 18 months to put an autonomous drone program together, he said. This lengthy start-up process can be intimidating to some industrial companies who are unfamiliar with the intricacies of running a safe and successful drone operation.

“That is not their business, right? They have a refinery to run or they have oil rig to run. they have something else that is their primary business and they are just using drones to help them run their primary business,” Gupta said.

“What we’ve realized is that we need to fix the business model. If we expect every organization to learn so much about drones before they can actually use this technology, it is not going to scale to what is required,” he said. Instead, companies should treat the adoption of drone system technology in the same way that they view security operations.

“In most organizations, security today is managed by security service providers,” he said. In a similar way, third-party drone system operators can tailor-make an autonomous UAV system to an individual industrial company’s needs.

“Those service providers are getting more and more sophisticated where they already have nationwide networks to where they can operate pretty much anywhere in the country,” Gupta said. “They already are investing in building remote operation centers from where they can operate drones at any facility that you might have across the U.S. or even in Europe.”

Drone system providers lease out the hardware, provide the trained pilots, and have the required nationwide waivers, which are needed to allow the client company to get started with an autonomous drone program in a matter of a few weeks, rather than several months or even years to establish a program from scratch, he said.

Humans, AI work together

As these autonomous drone systems begin to learn more, they can become more and more predictive instead of just being able to react to situations. “We are no longer running an inspection or doing other things manually, but these systems are able to understand and figure out what needs to be done in order to make sure that our facilities are up and running,” Gupta said.

At the same time, their human operators need to learn to work cooperatively with their systems to achieve the best outcome for their enterprises. “They are the inspectors,” he said. “They have to copilot with the application to where they are able to talk to this application and get the help from the AI in order to get the job done.”

Industrial operators are just beginning to get an idea of the potential value of AI-run systems, particularly with the advent of multimodal AI, which is able to process different modalities of data such as text, images and audio.

“It is able to understand voice, it is able to understand your historic record. It is able to understand your images videos from the past, so it has a full understanding of your assets,” Gupta said. “And whenever there is a failure or an issue, you are able to interact with AI and get a proper diagnosis done very fast.

This is the future that we are heading towards, where we make AI your coworker or copilot.”

Read more:

• FlytBase and Esri Partner to Streamline Drone Data into GIS Workflows
• How AI Is Pushing Industrial Drone Use Into a New Era
• Navigating the Future of the Drone Industry: Autonomy, AI, and Workforce Transformation

Jim Magill is a Houston-based writer with almost a quarter-century of experience covering technical and economic developments in the oil and gas industry. After retiring in December 2019 as a senior editor with S&P Global Platts, Jim began writing about emerging technologies, such as artificial intelligence, robots and drones, and the ways in which they’re contributing to our society. In addition to DroneLife, Jim is a contributor to Forbes.com and his work has appeared in the Houston Chronicle, U.S. News & World Report, and Unmanned Systems, a publication of the Association for Unmanned Vehicle Systems International.