DARPA Challenges Drive Dual-Use Innovation

Open competition with substantial prize-based incentives is a tool used by DARPA since 2004 to promote innovative solutions to national security problems.

30 April 2019
Stephen G. Anderson

Open competition with substantial prize-based incentives is a tool used by DARPA since 2004 to promote innovative solutions to national security problems. The original competition, DARPA's Grand Challenge for Autonomous Vehicles (2004, 2005, 2007), grew out of the United States Defense Authorization Bill for fiscal year 2001, which codified the goal of fielding unmanned ground combat vehicles for the US Armed Forces. In order to get US military personnel out of harm's way, that bill specified that one third of operational ground combat vehicles should be unmanned by 2015.

While that specific goal may have proven elusive, the process of open competition, which entails incentivizing the wider community to come up with a solution to the stated problem, was hugely successful according to Dr. Timothy Grayson, director of the Strategic Technology Office at DARPA and one of the judges for the original competition.

"I think the original Grand Challenge and the sequence of the self-driving challenges after that are a good example of why DARPA is doing the challenges," he said. "What I witnessed first-hand on those original self-driving car Grand Challenges, there was so much advancement in such a short time, it was almost mind-boggling."

"What that Grand Challenge pointed to was, if you just put the problem out there, and then unleash the innovation and creativity across all communities, some people will go home and work in their garages at night, absolutely titillated and motivated by the freedom to just explore on their own. It's all about immersion, experience, and trying things. That unfettered creativity is the huge benefit of the Grand Challenges."

The Challenges can be a boon for industry, too. While early ideas for a self-driving car date back to the 1920s, and Mercedes road-tested a self-driving van in 1986, the DARPA Challenge initiated significant development activity on multiple fronts and has spawned an industry that is forecast to grow at a compound annual rate of almost 40% to reach $557 billion in 2026.

Many of the newer commercial entities in this field, like Waymo and Uber, have links directly back to the DARPA Challenge, and there are now more than 20 commercial enterprises globally, including the traditional automakers, developing autonomous vehicles. New technology providers also have emerged as a direct outcome of the quest for a self-driving vehicle. These include many that offer photonics-based sensing systems, like the SPIE Prism Award-winning LiDAR companies Luminar and Blackmore.

The challenge of Challenges

Despite the undoubted benefits of these Challenges in terms of producing innovative technology solutions in a relatively short time, translation of these solutions into the national security world is anything but straightforward. By their nature, dual-use solutions developed commercially, while perhaps developed with a military problem in mind, are generally unaligned with a specific program or with the integration needs of a specific military system. As such, it can be extremely difficult for the Department of Defense (DoD) to incorporate them into its operational infrastructure.

Consider the application that motivated the original Challenge, robotic logistic convoys. What are all of the other things that a robotic logistics truck would have to plug into? At the very least, they would include several different command and control systems, each with different architectures. "Even if we had a perfect autonomous self-driving vehicle, you have a huge challenge integrating that with the rest of what you might call a system of systems architecture," noted Grayson, "and oh, by the way, there are humans involved too. How are the soldiers that are in the convoy going to be trained to interact with these self-driving vehicles?"

Monolithic systems and rigid architectures that take decades to develop characterize much of today's defense infrastructure. The challenge of the Challenges, so to speak, is how can the DoD take full advantage of rapidly developing technology solutions from the commercial world-solutions that may be appearing at rates that are orders of magnitude faster than the DoD is set up to deal with?

According to Grayson, DARPA is looking at this, considering ways to allow new capabilities to be ingested in smaller "baby steps," enabling faster uptake but with appropriate risk management. Typically, the reliability and security of a novel system must be proven up front, but what if continuous monitoring could ensure that any legacy capabilities remain uncompromised as the new capability comes online? Grayson cites the example of a new camera with an image format different from what the DoD legacy system expects. A software tool can automate translation between the data formats, but they also need a way to monitor how the new camera data is affecting the legacy platform.

Speed to launch

Meanwhile, the DARPA Challenges continue to drive innovation. On 18 April last year at the 34th Space Symposium in Colorado Springs, Colorado, DARPA formally announced its Launch Challenge. Competitors will vie for a top prize of more than $10 million based on their performance in two separate launches of small satellites at different launch sites on short notice. The idea is to incentivize industry to deliver launch systems with the capability to be both flexible and responsive, anticipating emerging DoD needs and changing the paradigm from years of launch preplanning and preparation to days.

Short launch cycles could also pave the way for rapid deployment of small satellites carrying imaging systems for things like disaster response (search and rescue) or severe weather observation. Rapid turnaround would likely also enable testing and iteration of new technologies, such as imaging systems or chips, and communications hardware prior to deployment on larger and more critical full-fledged satellite missions.

Today, it can take a decade to build, test, and launch a spacecraft. If the Launch Challenge is successful, decades may become days. While that may seem like an impossible goal, the success of the original Grand Challenge and subsequent autonomous vehicle challenges would suggest that "...unfettered creativity... across multiple communities" has a strong chance of successfully reaching that goal.

DARPA Grand Challenge


Enjoy this article?
Get similar news in your inbox
Get more stories from SPIE
Recent News
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?