CNA continuously invests in innovative, independent research projects that explore new tools and approaches for addressing emerging national safety and security challenges. These projects are showcased in the CNA Innovation Incubator (CNAi2). From analyzing machine learning for public safety to developing a Navy Force Design Lab, CNA's most creative thinkers are continuously working on new approaches to help government solve the nation’s toughest problems. In our Meet the Innovator series, we interview the analysts behind CNAi2 projects about their work and their innovation process.
Q: Your team won the top prize in a challenge sponsored by the National Institute of Standards and Technology. The challenge asked competing teams to use artificial intelligence to process sensor data from a natural disaster. What was innovative about your solution?
Crissman: This entire challenge was innovative. For the final competition, we were in Disaster City, part of Texas A&M’s Engineering Extension Service. Disaster City is a training center where first responders from all over the world go to train. They simulate buildings being destroyed, huge fires—and they have lots of sensors. We were in the command center, with each team in its corner, and they sent the teams unidentified streams of data from the sensors, from the Internet of Things. The machine learning algorithm we had built for the competition, which we call First Responder Awareness Monitoring during Emergencies, or FRAME™, had to determine: Was a particular set of data giving the wind speed, carbon monoxide levels, something else? It was like a hackathon.
The baseline requirement was to produce an Excel file for the judges to show how our solution had categorized and analyzed the data. But what really made our concept innovative is that we decided to visualize the data on a tool that first responders are currently using, to make the data clear and actionable. CivTAK is a geospatial mapping tool that lots of first responders already use on their phones or laptops. FRAME™ can take a wide variety of data from sensors and present it to emergency management unit commanders within CivTAK maps. Perhaps they would see an alert on a map showing a methane sensor detected a gas leak, and they could click on that sensor to see how dangerous the gas levels are. The judges really liked that our solution was first-responder-centric, focused on their situational awareness.
We also did a lot of thinking outside the box about where this could go in the future. FRAME™ could be used for more than sensors on streets and buildings. With wearables on first responders, FRAME™ could detect anomalies and signal, “Hey, this first responder could be having a health crisis.” It could perform sentiment analysis on social media data, using AI to analyze emotions expressed in posts related to an emergency and identify areas with high distress or potential threats. FRAME™ got off to a good start thanks to CNA funding, but with further government funding, there are a lot of other features that we could develop.
Q: How would you describe your personal approach to innovation?
Crissman: I never want to try to reinvent the wheel. If you want be innovative, you first have to explore the landscape and see what solutions are already out there, to learn from your predecessors. Then I see what I can build off of that. When I start to address a problem, I really try to become an expert in that area. It helps that I’m naturally curious. I strongly believe in the concept of “standing on the shoulders of giants.”
Q: You’ve worked as a computer scientist in academia, for the Army, and at CNA. How would you describe CNA's approach to innovation and research?
Crissman: CNA believes very strongly that data must come from reputable sources. And the data must not be biased. It’s better not to be too dependent on just one source of data. I’m involved in a CNA project to help the FAA develop an AI certification framework. In the research I’m doing on AI, we're looking at government. We're looking at industry. We're also looking at academia. Our framework will be more effective if we first understand how all of the different sectors are working with AI and what their best practices are.
CNA also stresses the importance of understanding those who ultimately use our research and analysis. In the case of FRAME™, that means understanding emergency management leaders and operators—something CNA has a lot of experience in. In the case of work for the FAA, I didn’t come into this knowing anything about air traffic controllers. But we’ve been meeting with air traffic control experts. Getting to know different stakeholders means we can create solutions that are more intuitive and easily usable.
Q: What inspires you to take on these challenges?
Crissman: I’m just one man, but I think it’s important to move the needle with my research in any way I can to improve safety, whether it’s during emergencies or in the national airspace. I want the airspace to be safe. The potential for even one preventable aviation incident motivates me. I aim to contribute my AI/ML expertise to enhance safety, from improving emergency response to optimizing airspace management. My goal is to make a tangible difference and ensure our skies and streets remain safe.
