Multi-mode Age of Information
When the U.S. military goes into battle, thousands of networked devices collect and transmit information to one another to track and adapt to the evolving landscape of the conflict. With all this information, it’s critical for devices to decide which data to collect, transmit, and process first.
The U.S. Department of Defense awarded a Multidisciplinary University Research Initiative (MURI) grant to a team of researchers, including four Virginia Tech faculty members, to study the fundamentals of latency in large-scale systems by developing a multi-mode age of information (AoI) concept to help a military internet of things (IoT) prioritize, process, and transmit data quickly and accurately in a highly dynamic environment. Jeffrey Reed, Willis G. Worcester Professor of Electrical and Computer Engineering, is the leader of the team.
AoI research has typically focused on timestamps: devices queue information based on when it is collected or received, but AoI helps them understand the age (or freshness) of the most recent information currently stored in the queue. The Virginia Tech team proposes a novel multi-mode AoI that will allow devices to also take into account the urgency and value of information when making these decisions. “We are developing a new theoretical foundation for optimizing age of information,” explains Tom Hou, Bradley Distinguished Professor of Electrical and Computer Engineering, “that will enable a large military internet of things to autonomously optimize latency and throughput in real time using multi-dimensional metrics.”
A military IoT includes thousands of devices, each with their own capabilities and limitations, spread across a large geographical area with changing network conditions—any activity could be suddenly interrupted or accelerated.
To apply their multi-mode AoI to a military IoT, the team must adapt it for use by many different types of equipment. Each device needs to be able to autonomously evaluate the cost and value of collecting and transmitting data based on its mission, taking into account disruptions, errors, and the complexity of measuring, processing, and using information.
“Our system has to be able to sustain extreme, but rare events, such as a sudden surge of traffic or a sudden drop in the quality of the wireless link, while maintaining its low latency,” says Reed. “We are developing very low complexity algorithms that can minimize the communication latency while operating with minimal or almost no overhead or coordination among massive devices.”
Consumer and Industry Applications
The new tools the team creates for optimizing latency in a massive military IoT could be ported to other wireless communications.
“I expect that our research is going to have a major impact on 5G communications,” explains Walid Saad, professor of ECE. “The distinguishing feature of 5G is the ability to process wireless applications quickly. It has low latency, and this property is especially important in new wireless applications such as vehicular control, augmented and virtual reality, tactile internet, manufacturing, and robotics in addition to military weapon systems.”
As the military IoT expands to include more devices, it will create new opportunities and motivations for cyberattacks. Many of the devices connected into the system, operating under weight, power, and processing constraints, will be vulnerable to hacking. “In a volatile military situation, physical attacks could also damage information security,” says Wenjing Lou, W.C. English Endowed Professor of Computer Science. “We will identify how devices and algorithms might be corrupted, jammed, or disrupted and close these vulnerabilities to ensure end-to-end information security throughout the military IoT.”