Johns Hopkins APL’s ‘Oz’ Reveals the Impact of Disaster Damage on Missions

Mon, 03/14/2022 - 13:16
Ajai Raj

Disasters are characterized by chaos — a disruption of the orderly flow of information that amplifies threats to life and property, hobbling effective countermeasures in situations where every second counts. Now, scientists and engineers at the Johns Hopkins Applied Physics Laboratory (APL) in Laurel, Maryland, have created a capability that can arm responders with crucial information by computing the impact of disaster damage on a given mission.

Dubbed “Oz,” the capability identifies impacted critical infrastructure and tracks dependencies and cascading effects during and after a disaster event. It leverages two tools previously developed at APL: Humanitarian Assistance and Disaster Relief (HADR) and Dagger. The Oz project team is currently demonstrating the platform to potential users across the federal government.

HADR uses artificial intelligence and machine learning (AI/ML) to scan overhead imagery and detect building damage and areas of flooding. Dagger tracks dependencies across a network of assets — people, facilities, systems and processes — and provides the user with a high-level awareness of the status of those assets and the missions and capabilities they support.

Both HADR and Dagger are operational capabilities, and government organizations are currently using Dagger to aid decision-makers during disasters.

Oz provides a user interface (UI) that allows the user to upload HADR output files and a critical infrastructure asset list; to compute risk assessments for each asset; and finally to output alerts and indicators to Dagger. If HADR is a set of eyes, and Dagger a brain, Oz serves as a kind of nervous system connecting the two. The result is a capability that can help responders adapt nimbly in an unfolding crisis.

The team behind Oz has developed three working demonstrations, using publicly available asset lists, that highlight the utility of the capability. The first tracks the effects of damage to health care facilities during 2020’s Hurricane Delta, which caused significant damage to the Gulf Coast; the second is focused on damage to U.S. Marine Corps facilities during 2018’s Hurricane Florence, which severely affected the Carolinas; and the third tracks damage to health care, police and military infrastructure during 2021’s Hurricane Ida, one of the most devastating and wide-ranging storms ever to make landfall in Louisiana.

These demos illustrate the utility of Oz in a disaster. It is able to ascertain and communicate damage to infrastructure, and the operational impact of that damage, as it occurs, said Melanie Lockhart, a project manager in APL’s Asymmetric Operations Sector (AOS) who helped manage HADR. Lockhart is leading the development of Oz, working on a team that also includes Greta Kintzley, Tony Webber and Jackie Soenneker.

“To use an example from Hurricane Ida, HADR reads overhead imagery to show that a police department was severely damaged by flooding,” said Lockhart. “Oz then feeds that into Dagger, which is able to put that into context from a mission perspective by showing that the security aspect of the mission is compromised, and to what degree. So you can see how a failure at the component level — whether that’s a police station that’s down, loss of personal protective equipment or other resources, or a hospital or urgent care clinic that is no longer available — can have a downstream impact on a particular aspect of mission readiness.”

HADR and Dagger represent considerable improvements over the previous state of the art in their respective use cases. Before HADR, analysts manually pored over imagery of damage from disasters, a painstaking and time-consuming process that is prone to error and does not permit fast dissemination of the resulting information. Dagger’s ability to track dependencies across diverse assets and deliver a continually updated assessment of the mission impact of damage to those assets — in a simple visual format that requires no special technical knowledge — is crucial for rapid decision-making. In bringing these tools together, Oz is a capability that is greater than the sum of its parts.

“Prior to Oz, we had HADR and Dagger, but there was no way for them to talk to each other,” Lockhart explained. “Now, we have a system where the user can set up their configuration in a simple UI, and they don’t have to touch it again.”

“So to return to the example of a hurricane, the user can set up their risk configurations in Oz, and Oz will process storm imagery as soon as it’s available,” she said. “Building damage and flooding are detected by HADR and risk assessments are fed into Dagger, which displays the operational impact, and all of this would be transparent to the user, in a near-real-time workflow.”

Lockhart also noted that Oz can be customized for a given use case to reflect the risk tolerance and assets of concern for any given user.

Andy Oak, mission area executive for Homeland Protection at APL, praised the ingenuity of the Oz team and the potential mission impact of the capability.

“Oz truly encapsulates what the Lab is all about — solving hard technical problems to create bold, innovative solutions that help make our nation a safer place,” Oak said. “Like the HADR and Dagger capabilities that Oz leverages, I am confident that Oz will empower leaders to make critical decisions in times of crisis.”