⚡ electrical engineering

Integrating Sensing into Covert Communications: Opportunities and Challenges

This article explores sensing-empowered covert communications, which leverage environmental and adversary information to transition from passive concealment to state-aware decision-making, thereby enhancing transmission reliability in dynamic channels while addressing new challenges related to exposure and resource consumption.

Original authors: Jun Wu, Xiaoqi Zhang, Haoyuan Pan, Gaosheng Zhao, Dong In Kim, Tse-Tin Chan

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Jun Wu, Xiaoqi Zhang, Haoyuan Pan, Gaosheng Zhao, Dong In Kim, Tse-Tin Chan

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to send a secret note to a friend in a crowded room, but a strict security guard (let's call him "Willie") is watching everyone. If Willie sees you whispering, he knows a secret is being passed. Your goal isn't just to make the note hard to read (which is standard security); your goal is to make it look like nothing is happening at all.

This paper explores a new way to pull off this trick: using "sensing" (like radar or cameras) to help you hide.

Here is the breakdown of the paper's ideas using simple analogies:

The Old Way: Guessing in the Dark

Traditionally, people tried to hide their messages by hoping the security guard was confused. They would shout a little bit of "white noise" (static) to mask their voice, hoping the guard couldn't tell if the voice was there or just part of the background noise.

The Problem: This is like trying to hide in a foggy room. If the guard moves around or the fog clears up, your "static" might not work. You might shout too loud and get caught, or shout too quiet and your friend can't hear you. You are essentially guessing blindly.

The New Way: Using a Flashlight to Hide

This paper suggests that instead of guessing, you should use a flashlight (sensing) to see where the guard is, what he is looking at, and where the shadows are.

The Catch: Shining a flashlight might actually make you more visible. If you turn on a bright light to see the guard, he sees the light!
The Solution: The paper argues that if you use the flashlight smartly, you can actually hide better. You can aim your secret message through a shadow you found, or use the flashlight beam itself to confuse the guard.

Three Ways to Use the Flashlight

The paper describes three main "modes" of doing this:

Look, Then Talk (Sensing-Then-Transmission):
- Analogy: You peek around the corner to see where the guard is, turn off the light, and then whisper your message.
- Pros: You know exactly where to hide.
- Cons: By the time you finish looking and start talking, the guard might have moved. Also, the time spent looking is time you aren't talking.
Talk and Look at the Same Time (Integrated Sensing):
- Analogy: You are holding a flashlight that also acts as a megaphone. You shine the light at the guard to blind him with glare while whispering your message to your friend.
- Pros: You don't lose time switching between looking and talking. The glare from the flashlight helps mask your whisper.
- Cons: It's a delicate balance. If the flashlight is too bright, the guard sees you. If it's too dim, you can't see where he is, and the glare isn't strong enough to hide you.
The Friend Helps You (Receiver-Assisted):
- Analogy: Instead of you holding the flashlight, your friend (Bob) holds it. He shines a bright, confusing light at the guard while you whisper. Since your friend knows the pattern of the light, he can filter it out and hear you. The guard just sees a blinding mess.
- Pros: You (the sender) stay in the dark and low-power, making you harder to spot.
- Cons: Your friend's equipment has to be perfect, or the "confusing light" might still leak some of your voice.

The "Smart" Flashlight (Intelligent Sensing)

The paper argues that you don't need a super-bright, high-definition flashlight that scans the whole room. That wastes energy and draws too much attention. Instead, you need an Intelligent Flashlight:

Multi-Modal: Instead of just using a radar, you also look at a map or use a camera. If the radar is blocked by a wall, the map tells you the guard is likely behind it.
Semantic (Meaning-Based): You don't need to know the exact distance to every chair in the room. You just need to know: "Is the guard looking at me?" or "Is there a shadow I can use?" You skip the details and focus only on what matters for hiding.
Predictive: You don't just look at where the guard is now; you guess where he will be in 5 seconds. This lets you prepare your hiding spot before he even gets there.
Cooperative: You and your friends share what they see. If one person can't see the guard because of a wall, another friend might see him, and they tell you.

The Real-World Test (The Drone Case Study)

The authors tested this idea using a drone (UAV) flying over a city.

The Scenario: A drone needs to send a secret message to a person on the ground while avoiding a security drone watching from the air.
The Result:
- Old Way: The drone scanned the whole sky randomly. It wasted energy and often got caught because it didn't know where the security drone was.
- New Way (Smart Sensing): The drone used "semantic sensing." It only looked for the specific "danger zones" (where the security drone was likely to be). It predicted where the security drone would move next.
- Outcome: The drone could send its secret message more reliably and with less chance of being detected, because it wasn't wasting energy scanning empty sky.

The Bottom Line

The paper concludes that sensing is a double-edged sword.

If you use it blindly, it exposes you.
If you use it intelligently (focusing only on what matters, predicting the future, and sharing info), it becomes a powerful tool to hide your existence.

The goal is no longer just to "hide in the noise," but to actively manage the environment so that your secret transmission looks exactly like nothing is happening at all.

Technical Summary: Integrating Sensing into Covert Communications

1. Problem Statement

Covert communications aim to hide the physical existence of wireless transmissions from unauthorized monitors (Wardens, denoted as Willie), rather than merely preventing unauthorized decoding. Conventional designs rely on exploiting uncertainty at the monitor—such as noise uncertainty or passive environmental interference—to mask the transmission. However, these "blind" approaches become ineffective in dynamic propagation environments where the monitor's state (location, channel conditions) varies rapidly. Without reliable knowledge of the Warden's channel, friendly jamming is often generated without spatial guidance, consuming excessive power and potentially impairing the legitimate receiver (Bob) while offering only marginal improvements in covertness.

The core challenge addressed is the coupling between reliable transmission and low detectability: stronger signals improve the legitimate link but increase detection risk. Furthermore, simply adding active sensing introduces new risks, as sensing emissions can inadvertently expose the transmitter and consume resources (time, power) needed for data transmission. The paper posits that sensing must be carefully integrated to shift covert design from passive concealment to state-aware decision-making without compromising the covertness constraint.

2. Methodology and Architectures

The paper investigates "sensing-empowered covert communications," where adversary and environmental information guide transmission and jamming control. It defines the monitoring process as a binary hypothesis test where the goal is to make the observation under transmission ( $H_1$ ) statistically indistinguishable from silence ( $H_0$ ), even when auxiliary sensing or jamming waveforms are present.

The authors propose and analyze three representative architectures (illustrated in Fig. 1 of the paper):

Sensing-Then-Covert Transmission: A sequential approach where external sensing nodes characterize the environment before data transmission. A fusion center processes this data to estimate Willie's location or blockage state, allowing Alice (the transmitter) to adapt her policy. This enables active adversary-aware design but suffers from latency and the risk of outdated state information in mobile scenarios.
Integrated Sensing and Jamming (ISAC-based): This mode utilizes a unified radio platform where sensing and communication share waveforms, spectrum, and hardware. The sensing waveform is designed not just for estimation but as a "cover signal" or friendly interference. By randomizing waveform statistics and decoupling scheduling from the covert-data state, the system masks Alice's signal leakage at Willie. This approach offers spatial focusing of interference but creates a tight coupling between sensing power, estimation accuracy, and exposure risk.
Full-Duplex Receiver-Assisted Covert Transmission: This architecture shifts the environmental shaping burden to the full-duplex receiver (Bob). Bob broadcasts a dual-purpose probing-jamming waveform. From Willie's perspective, this acts as artificial noise; from Alice's perspective, it provides channel state information. Bob suppresses self-interference to recover the message. This reduces the transmitter's direct exposure but is limited by hardware impairments and residual self-interference.

3. Intelligent Sensing Paradigms

To address the trade-off between sensing accuracy and exposure, the paper outlines four intelligent sensing paradigms designed to extract task-relevant information under limited probing budgets:

Multi-Modal Sensing: Fuses heterogeneous data sources (e.g., visual observations, historical mobility data, environmental maps) to infer the adversary's state. This reduces reliance on active radio probing and improves robustness against blockages and non-line-of-sight (NLoS) conditions.
Semantic Sensing: Shifts the objective from reconstructing physical parameters (e.g., exact coordinates) to extracting compact, task-relevant variables (e.g., risk indicators, threat levels). This reduces signaling overhead and exposure by avoiding unnecessary high-dimensional channel reconstruction.
Predictive Sensing: Utilizes historical observations to forecast the future evolution of Willie's mobility and channel conditions. This enables proactive risk management and beamforming adjustments before transient vulnerabilities occur, addressing the latency issues of reactive sensing.
Cooperative Sensing: Forms a distributed perception network where multiple legitimate nodes share observations (raw data, features, or semantics) to mitigate sensing blind spots and enhance detection in occluded regions.

4. Case Study: Low-Altitude Wireless Networks (LAWNs)

The paper presents a case study in urban Low-Altitude Wireless Networks (LAWNs) involving an ISAC-enabled UAV (Alice) communicating with a ground user (Bob) while avoiding an aerial Warden (Willie).

Setup: The scenario operates in the millimeter-wave band with rapid LoS/NLoS transitions. The study compares three schemes: Conventional Sensing (broad scanning), Predictive Beamforming, and Semantic Beamforming.
Findings:
- Conventional Sensing: Performs broad spatial scanning, leading to the degradation of sensing effectiveness as communication rate requirements increase. It wastes resources on directions irrelevant to covertness.
- Predictive Beamforming: Exploits historical data to focus on likely Warden regions, improving the trade-off but remaining susceptible to mobility-induced mismatches.
- Semantic Beamforming: Focuses sensing energy specifically on covertness-relevant features (e.g., leakage-dominant directions) rather than full environmental reconstruction.
Results: The study demonstrates that semantic beamforming achieves a more stable performance trend and better spatial resource utilization. It effectively reallocates sensing attention to environmental features most relevant to the covert transmission, reducing unnecessary exposure while maintaining communication reliability.

5. Key Contributions and Claims

The paper claims the following contributions:

Architectural Framework: It presents representative architectures for sensing-empowered covert communications, highlighting how sensing transforms design from blind uncertainty exploitation to state-aware decision-making. It identifies new design considerations, including composite performance metrics (combining covertness and sensing accuracy, e.g., CRLB), expanded optimization degrees of freedom (DoFs), and coupled constraints.
Intelligent Sensing Paradigms: It discusses novel paradigms (Multi-modal, Semantic, Predictive, Cooperative) that prioritize decision-relevant information over maximum sensing accuracy to minimize exposure.
Trade-off Analysis: The paper elucidates the fundamental trade-off where accurate sensing improves adversary awareness but consumes resources that could be used for data or jamming, and conversely, allocating more resources to data may increase detection risk.
Future Directions: It outlines promising research directions, including AI-enabled covert intelligence for adaptive resource allocation, spectrum-map-aided networking for identifying low-risk regions, and quantum-based covert communications for hardware-based randomness and secure coordination.

6. Significance

The paper argues that sensing-empowered covert communications represent a necessary evolution for dynamic wireless environments. By moving away from static, blind policies toward adaptive, state-aware designs, legitimate systems can better manage the exposure risk. However, the paper emphasizes that this integration is not trivial; it introduces a new design problem requiring the joint optimization of sensing accuracy, communication reliability, and covertness. The significance lies in demonstrating that with intelligent sensing paradigms, it is possible to enhance spatial resource utilization and the reliability of covert data delivery in time-varying channels without incurring prohibitive exposure costs.