As unmanned aerial vehicles and satellites are used more and more, understanding how they interact with everything else in the world is increasingly important.
Taylor Johnson, an assistant professor of Computer Science and Engineering at The University of Texas at Arlington, will investigate emergent behavior in cyber-physical systems such as UAVs and satellites using a two-year, $499,546 grant from the Air Force Research Laboratory to model, predict, monitor and control potential emergent behavior.
Emergent properties are those that spontaneously emerge in the execution of engineered systems, particularly when those systems have software and physical components, such as unmanned aerial vehicles or satellites. For example, a group of unmanned aerial vehicles may be programmed to fly in close proximity formations while avoiding collisions. That group may form into a flock using local commands to avoid contact, without an actual command to form a flock specified anywhere.
"At present, we can't predict how a cyber-physical system will perform at the global scale. Emergence could cause a system to perform in unexpected ways, which could lead to unsafe scenarios or performance degradation," Johnson said. "We need methods to determine for a given system, if emergent behaviors are possible, could an attacker exploit these emergent behaviors at the software level or insert physical signals that would lead to global specifications being violated, such as drones colliding?"
Khosrow Behbehani, dean of the UT Arlington College of Engineering, said Johnson's work would contribute to knowledge that will make cyber-physical systems more stable and predictable.
"This work will contribute to our understanding of how cyber-physical systems may react in situations where they interact with each other, with humans or with objects," Behbehani said. "Future work in artificial intelligence, satellite navigation, and unmanned vehicles will benefit greatly from Dr. Johnson's findings."
Johnson's research will address four tasks:
- How to define emergent behavior rigorously using formal specifications;
- How to verify the existence or absence of emergent behavior with formal verification;
- How to control to achieve desired, or avoid undesired, emergent behavior;
- How to evaluate the findings in a case study.
Johnson will create formal specifications in a mathematical language to describe emergent behavior. From there, he will attempt to define what the existence, or lack of existence, of emergence means for a given formal model of a cyber-physical system.
Johnson will develop formal methods to verify the existence or absence of emergent behavior.
Johnson will find ways to maintain or prevent emergent behaviors using run-time assurance and self-stabilization, which are foundational techniques in runtime verification and distributed systems. Together, these methods are expected to steer the distributed cyber-physical system back to its intended state if it strays toward unintended emergent behavior.
Finally, Johnson will evaluate these findings by building software tools to perform the verification, and deploy the runtime methods in quadcopters. Running the software and the quadcopters concurrently will allow him to see emergence as it happens and test methods to avoid it if it is undesired. In the long-term, these methods may be effective for identifying more general emergent behavior, such as what many imagine in the future of artificial intelligence.
About UT Arlington
The University of Texas at Arlington is a comprehensive research institution of more than 50,000 students in campus-based and online degree programs and is the second largest institution in The University of Texas System. The Chronicle of Higher Education ranked UT Arlington as the seventh fastest-growing public research university in 2013. U.S. News & World Report ranks UT Arlington fifth in the nation for undergraduate diversity. Visit http://www.