Experimental Covert Communication Using Software-Defined Radio

Rohan Bali, Trevor E. Bailey, Michael S. Bullock, and Boulat A. Bash · 2025

Read Experimental Covert Communication Using Software-Defined Radio by Rohan Bali, Trevor E. Bailey, Michael S. Bullock, and Boulat A. Bash (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-02297.txt.

Introduction

Rohan Bali and co-authors show, in a 2025 paper, hid radio talk that can be proved safe in a lab build. Hid talk sends so faint that a watch foe can not tell it runs. A square root law caps it: hid bits grow with the root of the time-band span, no more. Past that cap, the talk is caught or read wrong. Lab proof was thin: two past builds, both on light. This team built the radio form with software radios. Their first reads match the law's calls.

The build shows the math holds on air, not just on glass. It opens a path to field rigs and names the next asks for them. Safe hid talk may suit more than spies.

Catalog record

CREATORRohan Bali, Trevor E. Bailey, Michael S. Bullock, and Boulat A. Bash
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-02297

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Experimental Covert Communication Using Software-Defined Radio Thanks: This work was supported, in part, by the National Science Foundation under Grants No. CCF-2006679 and CNS-2107265. Rohan Bali, Trevor E. Bailey, Michael S. Bullock, and Boulat A. Bash Abstract The fundamental information-theoretic limits of covert, or low probability of detection (LPD), communication have been extensively studied for over a decade, resulting in the square root law (SRL): only L ​ n L\sqrt{n} covert bits can be reliably transmitted over time-bandwidth product n n , for constant L > 0 L>0 . Transmitting more either results in detection or decoding errors. The SRL imposes significant constraints on hardware realization of provably-secure covert communication. Thus, experimental validation of covert communication is underexplored: to date, only two experimental studies of SRL-based covert communication are available, both focusing on optical channels. Here, we report our initial results demonstrating the provably-secure covert radio-frequency (RF) communication using software-defined radios (SDRs). These validate theoretical predictions, open practical avenues for implementing covert communication systems, as well as raise future research questions. I Introduction Warfighter operation in highly contested settings demands covert or low probability of detection/intercept (LPD/LPI) communication that enables message transmission without alerting an adversary [ 1 , 2 , 3 , 4 ] . This contrasts the traditional cryptographic [ 5 ] and information-theoretic secrecy [ 6 ] security that prevents access to the transmission’s content, but not its detection. Careful waveform design and spread spectrum techniques are often employed in practice to reduce adversaries’ signal-to-noise ratio (SNR) below the noise floor [ 7 , Pt. 1, Ch. 5] . However, guaranteeing covertness requires intricate mathematical analysis which yields the square root law (SRL): only B(n)=L\sqrt{n} covert bits can be reliably transmitted over n channel uses [ 1 , 2 , 3 , 4 ] . The channel-dependent constant L>0 is called covert capacity and n=TW is the transmission time-bandwidth product. Notably, the associated Shannon capacity [ 8 ] is zero, since \lim_{n\to\infty}\frac{B(n)}{n}=0 . This is because adversary in covert communication seeks just one bit of information (whether transmitter is on or not) versus \mathcal{O}(n) bits of transmitted data in traditional secure communication. Nevertheless, a significant number of such provably-covert bits can still be transmitted. The discovery of the SRL in [ 1 , 2 ] resulted in an explosion of research by the communication and information theory communities overviewed in a tutorial [ 3 ] and a detailed survey [ 4 ] . This includes characterization of capacity L for additive white Gaussian noise (AWGN) and discrete memoryless channels (DMCs) [ 9 , 10 , 11 , 12 ] , covert networks [ 13 , 14 , 15 , 16 , 17 , 18 ] , quantum aspects of covert communication [ 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 ] , and many other directions. Notwithstanding this progress on fundamental theory, experimental SRL-based covert communication remains underexplored, with only two published works, both focusing on optical channels [ 20 , 31 ] . This paper addresses this gap by providing the first, to our knowledge, experimental validation of the SRL on radio frequency (RF) channels. We evaluate a covert communication protocol using software-defined radio (SDR). Our implementation uses USRP X310 SDR units deployed on the ORBIT testbed [ 32 , 33 ] , as detailed in Section III-A , enabling controlled and reproducible experiments. Binary phase-shift keying (BPSK) with a Gaussian pulse shaping filter allows transmission with controlled temporal and spectral symbol leakage while mitigating timing jitter. To meet the SRL, the transmitter uses a “sparse coding” strategy, selecting a random subset of available channel uses that is secretly shared with the receiver in advance. An experimental framework based on a synthetic Gaussian noise source within a wired, shielded network provides environment control and supports reproducibility. The SRL governs, arguably, the worst case scenario. As detailed in Section II , it assumes that the characteristics of the transmitter-adversary channel (noise power variance and transmission loss), the exact time and frequency band of potential transmission, and details of the transceiver system design are known to the adversary. However, the adversary 1) cannot control all the random channel noise; and, 2) lacks access to the secret shared between transmitter and receiver prior to the transmission. We implement this model and show that one can transmit covertly under such conditions. While conservative, it provides a high level of security to unforeseen adversarial technological surprises. Indeed, relaxing these assumptions can result in significant performance gains. For example, an adversary’s uncertainty of transmission time/frequency yields a multiplicative improvement to covert capacity L [ 34 , 35 , 36 ] , while uncertainty in noise power level can lead to a linear law: \mathcal{O}(n) covert bits reliably transmissible over n channel uses [ 37 , 38 ] . We defer validation of these to future work. Additionally, our experiment motivates a systematic study of the impact of adversary hardware limitations, such as sampling timing jitter and receiver bandwidth constraints. Furthermore, employing more efficient modulation and coding schemes (such as [ 39 ] ), as well as minimizing the pre-shared secret size (see [ 9 ] ) are avenues for future exploration. Fig. 1: Covert communication over an AWGN channel. Alice ( a ) transmits a complex-valued covert message \vec{u}\in\mathbb{C}^{n} over n uses of a channel corrupted by independent AWGN at intended receiver Bob ( b ) and adversary (warden) Willie ( w ). Here \vec{u} is a sequence of pulse-shaped symbols and empty pulse slots described in Section II-B . Bob receives h_{a,b}e^{j\theta_{a,b}}\vec{u}+\vec{z}^{(b)} while Willie observes h_{a,w}e^{j\theta_{a,w}}\vec{u}+\vec{z}^{(w)} . For r\in\{b,w\} , path loss h_{a,r} and phase \theta_{a,r} are static and known, while \vec{z}^{(r)} is complex circularly-symmetric AWGN. Alice and Bob’s pre-shared secret allows reliable message decoding while rendering it indistinguishable from noise \vec{z}^{(w)} by Willie. This paper is organized as follows: next we develop the theoretical foundation for RF covert communication by deriving a sparse-coded BPSK transmission scheme that satisfies the SRL for the discrete-time AWGN channel model. Section III details the experiment on the ORBIT testbed, and presents our results. Finally, Section IV interprets these, highlighting practical design challenges and their implications for practical systems, and outlines future work. Appendices A - C provide supporting derivations. II Theoretical System Analysis II-A Channel Model Consider a static discrete-time AWGN channel model depicted and described in Fig. 1 . Per the square root law (SRL), one can transmit B(n)=L\sqrt{n} covert bits reliably in n uses of such channel [ 1 , 2 ] , with covert capacity L derived in [ 9 , 10 ] . We adapt the results in [ 1 , 2 ] to our SDR-based system model in the following subsections. II-B System Model and Reliable Covert Communication The AWGN channel described in Fig. 1 provides a discrete-time model of the covert communication scheme implemented in Section III-A . Analysis of the fundamental limits in such model [ 1 , 2 ] often treats each channel use separately and assumes that Alice can use arbitrarily low output power. However, practical radio systems require pulse-shaping for bandwidth efficiency, and to mitigate inter-symbol interference and timing

Read the complete text