LAS VEGAS -- Technologies like artificial intelligence and deep learning are driving the evolution of drones and fueling their autonomous future, according to Jesse Clayton, senior manager of product management for intelligent machines at Nvidia. Clayton spoke with SearchCIO at the recent InterDrone conference in Las Vegas, where he discussed the underlying technologies that are shaping the future of drones. In this video, he gives an overview of the commercial applications of drones and explains how advances in AI are impacting the drone industry.
What are some of the most surprising business applications of drones that you have seen?
Jesse Clayton: Before we talk about the applications, it's important to understand some of the big trends that are happening in technology right now. Modern artificial intelligence -- and, specifically, deep learning -- is making it possible to solve problems that were practically unsolvable before. There are three big things that have happened to make this possible: The first is big data, the second is new algorithms, and the third is GPUs from Nvidia. Artificial intelligence is so powerful that it's impacting almost every industry.
There's tremendous opportunity to bring this artificial intelligence to robots like UAVs [unmanned aerial vehicles] to enable them to do new things, and that's why we built Nvidia Jetson. We've seen a lot of adoption from our customers that enable them to do new things. For example, a company called Intelligent Flying Machines built a drone that can autonomously navigate through a warehouse and match what's on the shelves to what's in the inventory system to help the distribution center manage inventory better.
A company called Skycatch uses UAV data and processes that, with artificial intelligence, help construction sites manage their sites more efficiently.
And then Aerialtronics is using drones for industrial inspection. They have a UAV that can autonomously fly around wind turbines and cell towers, detect faults and generate reports more quickly so those faults can be repaired. We see a lot of opportunity in other areas like precision agriculture, package delivery, safety and security, and search and rescue.
Is it more worthwhile for companies to invest in engineering hardware or software?
Clayton: We see some of the biggest innovation happening on the software side. There's so much research that's happening now for new neural network topologies and new techniques in artificial intelligence to solve different problems. We see a lot of potential to bring those capabilities and apply them to UAVs.
Let's talk about the future of drones. What's in store for enterprise drone software?
Clayton: There's a lot of research happening now on the AI side. We see this continuing to accelerate and we see companies applying this research to enable new applications in the UAV space.
What does the future of drones look like?
Clayton: Over the next five years, we see a transition to drones that are able to navigate and solve tasks more autonomously. For the areas that I mentioned -- industrial inspection, precision agriculture, package delivery, safety and security, search and rescue -- there's going to be an opportunity for UAVs to solve these challenges in a way they haven't been able to before.
What's Nvidia's game plan when it comes to drones?
Clayton: Nvidia makes Jetson, and Jetson is Nvidia's platform for artificial intelligence for edge devices like UAVs. Also, we just launched a new project called Redtail -- a framework for enabling autonomous navigation aboard UAVs. One of the challenges with UAV navigation is that, oftentimes, you may not have GPS, you're in a GPS-denied environment or the application may not be suitable for a pilot to do the task. For those applications, you need to have something that's more intelligent and more powerful. Redtail is a framework to bring in artificial intelligence to help drones navigate more effectively.