Is signal interference defense inadequate for the task?

In today's fast-paced world, high-speed Internet and mobile services have become necessities rather than luxuries. But as data services continue to evolve and mature, requiring higher Internet speeds, and as operating systems need protection, hackers and adversaries continue to interfere. For some, this involves hacking into wireless connections in homes and offices for the purpose of extracting personal or business data.

But whether they're targeting government agencies, private companies, or individual users, these attackers typically use high-power signal jamming devices - wireless portable devices that block communication between devices. Such jammers are also a means of defense for users seeking to escape these attacks.

With this dichotomy in mind, former Khoury doctoral student Hai Nguyen and his advisor Guevara Noubir, both members of the Institute for Cybersecurity and Privacy at Khoury College, have pioneered a novel approach that can essentially eliminate these high-performance cellphone jammer in the event that traditional techniques fail. This failure occurs because traditional technologies are designed for non-malicious interference, require mechanical moving parts that are slow to respond to interference, or require additional radio frequency bands to achieve resilience. Their technique, known as JaX, can circumvent these scenarios.

"When you use wireless communication, you want to make it as robust and reliable as possible. As researchers, we like to solve difficult problems, and this problem has always been there, "Nubier said. "We want to approach this problem from a unique perspective, which is to develop secure communication technologies for GPS and Wi-Fi."

Nguyen and Noubir started working on this machine learning-driven idea about a year ago, when signal jamming played a major role in the Russia-Ukraine war. As they built up large amounts of synthetic data and collected experimental materials, their goal was to build a kind of black box that could be placed in front of the network and seamlessly protect users. The solution does not require complex machinery, thus simplifying the process of eliminating high-power signal gps jammer.

"For data collection, we set up a test bench, and for transmitters, we used software radios, and for jammers and receivers, we used other radios," explains Nubier. "We sent a legitimate signal, and then we also sent interference. We receive them through two antennas that store data, so we know what we are sending and what is interfering.

"The cool thing about our work," he added, "is that we can get an adversary's jammer to emit a more powerful signal than a legitimate one, and we can eliminate the interference as if it never existed!"

Nubier believes the paper's findings could be useful for a wide range of individual and organizational users who want to avoid malicious attacks on their data.

"The fact that we were able to achieve a very low bit error rate shows that we can outperform an adversary that can transmit a hundred times faster than a legitimate signal," he said. "There can be an extremely noisy signal in the environment, but because we have JaX, we can estimate what the gsm jammer is really transmitting, subtract it, and just leave the legitimate signal."

While this process sounds promising, the benefits only matter if the technology is logically sound and affordable - two barriers that often prevent innovative technologies from becoming ubiquitous. But Noubier believes he and Nguyen have cleared that hurdle, too.

"Our machine learning model is compact, so if you're designing any new system, you can easily incorporate it," he says. "Again, if the Department of Defense wanted to integrate JaX, the cost would be negligible compared to other models, and the model would be much more efficient."

Noubir and Nguyen, who recently defended their dissertation and joined Meta as a research scientist, acknowledge that new approaches to old problems always need tweaking and fixing. Still, they expect JaX to solve long-standing network problems and want to bring it to market.