Breaking the square-root barrier in covert communications
Abstract
The focus of this project is on covert communications. In covert communications, there is a legitimate transmitter Alice, a legitimate receiver Bob, and a warden Willie. The goal of covert communications is that, Alice and Bob communicate, and keep it hidden from Willie, i.e., Willie should not be able to detect that a communication is taking place. Covert communications is even more restrictive than secure communications, in that, while in secure communications the goal is to secure the content of the message from the adversary, in covert communications the goal is to hide even the fact that a communication is taking place from the advisory. That is, in secure communications, an adversary knows that there is communication going on between Alice and Bob, but is unable to decode the messages; in contrast, in covert communications, the adversary cannot even tell that there is communication going on. This severe restriction has resulted in what is now known as the square-root law, which states that, for a communication to be covert, at most sqrt n bits can be transmitted in n channel uses. In essence, sqrt n is the limit to which information can be hidden in the noise. This means that the rate of communication per channel use, which is sqrt n divided by n, goes to zero as n increases. That is, the covert communication rate is zero. The goal of this exploratory STIR project is to develop methods and techniques such that non-zero, i.e., strictly positive, covert communication rates can be achieved. This project will consider two communication scenarios: 1) wireless communications using radio frequency (RF) signals, and 2) communications using the timings of the packets. The first setting brings us to the classical Gaussian channel model with or without fading and interference, and the second setting brings us to the timing channel model. In the RF domain, our main goal is to exploit the multi-user nature of the wireless channel and carefully design user interactions to enable a subset of a network of users to have covert communication with their receivers with the help of the remaining users. In the timing domain, even in a single-user channel, covert communication can be possible. In the timing channel, we will design covert communication schemes for the canonical single channel, and consider multi-user effects. The potential impact of the proposed project on Army s capabilities includes the development of robust covert communication methods in realistic communication scenarios. Such methods will bring significant advantages to Army s communication capabilities in contested and unfriendly environments. This is a collaborative project with Dr. Ananthram Swami of ARL. We will have frequent visits and meetings with Dr. Swami and a robust collaboration plan. The requested funds in this STIR project is S60K for 9 months. The requested funds will be used toward graduate research assistant (GRA) salary and PI time.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2018
- Source ID
- W911NF1710012
Entities
People
- Sennur Ulukus
Organizations
- Army Contracting Command
- United States Army
- University of Maryland