Drone racing league launches new AI competition

The Drone Racing League (DRL), a global, professional drone racing circuit unveiled their latest offering: A new autonomous drone series designed to test the limits of artificial intelligence (AI).

Since 2015, the DRL has sought to expand the capabilities of drones at a time when dozens of industries use them for a variety of purposes. DRL RacerAI will be the league’s first autonomous racing drone, and they hope that one day it could beat a drone piloted by a human.

“The DRL RacerAI is a historic moment for The Drone Racing League and the future of sports. AI has defeated humans in nearly every digital game we know, but it hasn’t come close to defeating a human in real-life sports — until now,” said DRL CEO and founder Nicholas Horbaczewski, in a press release. “Through the competitive AIRR events, we’ll watch the DRL RacerAI get faster and smarter, catch up to human competitors, and one day, outpace the best pilot in the world. This will mark an initial step towards a future when autonomous systems can relieve humans from performing often dangerous disaster-relief efforts and search and rescue missions.”

SEE: AI in healthcare: An insider’s guide (free PDF) (TechRepublic Premium)

DRL made the announcement ahead of their inaugural Artificial Intelligence Robotic Racing Circuit, which hosted its first event in Orlando and will have races throughout the fall in Washington DC, Baltimore, and Austin. 

The races are the culmination of a yearlong process involving more than 2,300 industry leaders on more than 420 teams from 81 countries. 

In September 2018, security company Lockheed Martin announced a partnership with DRL and the creation of the AlphaPilot Innovation Challenge, an open innovation challenge, developing artificial intelligence for high-speed racing drones.  

Thousands of university students, technologists, coders and drone enthusiasts joined the competition seeking to combine AI, machine learning and fully autonomous flight. The goal was to create an AI framework that could pilot racing drones through high-speed aerial courses without any GPS, data relay or human intervention.

“At Lockheed Martin, we are working to pioneer state-of-the-art, AI-enabled technologies that can help solve some of the world’s most complex challenges — from fighting wildfires and saving lives during natural disasters to exploring the farthest reaches of deep space,” said Lockheed Martin Chief Technology Officer Keoki Jackson. “Now, we are inviting the next generation of AI innovators to join us with our AlphaPilot Innovation Challenge. Competitors will have an opportunity to define the future of autonomy and AI and help our world leverage these promising technologies to build a brighter future.”

Lockheed Martin offered the winning team $1 million and an extra $250,000 if the winning AI drone could beat a human-piloted drone. The winning drone will race the fastest 2019 DRL Allianz World Champion pilot at the end of the season. 

After whittling down the field, DRL and Lockheed Martin announced the nine finalists who will compete in the four races in the coming months. Many of the teams hail from universities like UCLA and the University of Zurich or scientific institutions in countries like the Netherlands, South Korea, Switzerland, Poland, and Brazil.

Nearly 70 students, drone technologists, aerospace engineers, and coders are involved in the Artificial Intelligence Robotic Racing Circuit.

Ryan Gury, chief technology officer of the DRL, spoke to TechRepublic about the league’s desire to push the bounds of what the world thinks drones can do.  

“Robotics are becoming more and more physical. The day that they start to outpace us is becoming inevitable. We want to be part of that conversation, help foster advanced robotics and create innovative competition where we can see advancement in robotics beat a human in a sport,” he said.  

“My team had to create all the building blocks necessary to deliver that, including, the development environment where each team will code their autonomous strategy and math and algorithms,” Gury continued. “We had to develop a simulation so that they could do this from their desk, and it wouldn’t destroy a drone each time they made a mistake. We had to create a very advanced racing drone that’s using state of the art parts, well beyond what’s available on the consumer side.”

Gury added that the competition was a great way to motivate people to pushing beyond what they previously thought drones were capable of.  

DRL’s human-piloted drone league has races around the world in 3D racing environments like museums and arenas. The course is usually one mile long and each race has 12 pilots. The DRL Allianz World Championship Season drew six million viewers on Twitter in August and will have its broadcast premiere on NBC soon. 

Gury said they don’t expect the AI drone to beat a human-piloted one this year but that as the technology gets better, it will be harder to beat.

“I grew up in the early 80s. I love Nightrider, loved watching Space Odyssey and loved Watson on Jeopardy. I love computers, so I want to see what’s possible,” he said. “The idea of this thing is to really utilize competition to get to the point where this starts to eclipse humanity. To see the first time a machine beats a human. To see when machines can compete against humans. The evolution of sports is going to happen and this will be a big part of it.”

Also see