Resource Allocation for Secure Dual-UAV-Assisted ISAC System

summary

The gist

Integrated sensing and communication (ISAC) is a rising technology in the next wireless communication networks…

In short

The research optimizes a secure Integrated Sensing and Communication (ISAC) system using two aerial vehicles: a base station UAV and a jammer UAV. The system simultaneously communicates with users and senses potential eavesdroppers, while the jammer uses sensed location data to suppress the eavesdropper. It solves a complex non-convex problem through an iterative algorithm to maximize the average secrecy rate.

Key concepts

Integrated Sensing and Communication (ISAC)
ISAC combines communication (sending data) and sensing (detecting things like eavesdroppers) into one system using limited spectrum. This allows the UAV to perform both tasks at the same time, making it more efficient by sharing resources.
Dual-UAV-Assisted System
This setup uses two unmanned aerial vehicles: one acts as a base station for communication and sensing, and another acts as a friendly jammer. They work together to ensure secure communication by detecting threats and actively jamming them.
Secrecy Rate Maximization
This is the goal of the optimization: maximizing the rate at which information can be sent securely. Security is achieved by ensuring that the eavesdropper cannot decode the transmitted data, even though they are present in a potentially hostile environment.

Terminology used across episodes

This episode discusses

The paper

Resource Allocation for Secure Dual-UAV-Assisted ISAC System · Read on arXiv

Hongjiang Lei, Jianshuo Geng, Ki-Hong Park, Jia Ye, Liang Yang, Xiaqing Miao, Gaofeng Pan

Chongqing University of Posts and Telecommunications School of Communications and Information Engineering Department at Chongqing University of Posts and Telecommunications · King Abdullah University of Science and Technology CEMSE Division at King Abdullah University of Science and Technology · State Key Laboratory of Power Transmission Equipment Technology School of Electrical Engineering Department at Chongqing University · College of Computer Science and Electronic Engineering Hunan University School of Information Engineering Department at Hunan University · Beijing Institute of Technology School of Cyberspace Science and Technology Department at Beijing Institute of Technology

Integrated sensing and communication (ISAC) is a rising technology in the next wireless communication networks, enabling the simultaneous execution of communication and sensing tasks by fully utilizing limited spectrum resources. In this work, we investigate the secrecy performance of a dual-uncrewed aerial vehicle (UAV)-assisted secure ISAC system. Specifically, a base station UAV communicates with users and transmits radar signals to locate potential eavesdroppers, while simultaneously providing information to a jammer UAV to perform jamming tasks. Considering constraints such as maximum UAV velocity, transmit power, propulsion energy, and sensing thresholds, we maximize the average secrecy rate by optimizing user scheduling strategies, time allocation, transmit power, and UAV trajectories. The presence of a non-convex problem, originating from tightly coupled variables, is tackled by an efficient iterative algorithm. In particular, the original optimization problem is decomposed into six subproblems, and non-convex subproblems are transformed into approximately convex forms via successive convex approximation. Then, block coordinate descent techniques are employed to solve all subproblems sequentially. Numerical results demonstrate the convergence and effectiveness of the proposed algorithm.

DOI: 10.1016/j.dcan.2026.11.005

Transcript

Introduction to the show: ident: Computer Science. The day in computer science: software, systems and theory on arXiv.

Tom: We're looking at the paper: "Resource Allocation for Secure Dual-UAV-Assisted ISAC System".

Title and authors: Tom: So we’re looking at the paper "Resource Allocation for Secure Dual-UAV-Assisted ISAC System," and it’s tackling how you use limited spectrum resources when you need both communication and sensing at the same time.

Jane: Exactly, it's about this dual-uncrewed aerial vehicle system where one drone talks to users while sending radar signals to find eavesdroppers, and another drone helps jam those eavesdroppers.

Lu: The authors are looking at maximizing the secrecy rate under a bunch of real constraints like maximum UAV velocity, transmit power limits, propulsion energy usage, and sensing thresholds.

Meng: But they're not just throwing variables at it; they’re optimizing user scheduling strategies, time allocation ratios, transmit power levels, beamforming vectors for communication and sensing signals, and the actual flight trajectories of both drones.

Tom: And what makes this interesting is that the problem they formulate is non-convex because all these variables are tightly coupled together in a complex way.

Jane: So to solve this complexity, they propose an efficient iterative algorithm to break down that giant non-convex problem into smaller, more manageable pieces.

Lu: They use something called block coordinate descent to decompose the original problem into six subproblems: user scheduling, time allocation, transmit power of the base station UAV, beamforming optimization for sensing and communication signals, trajectory optimization for the base station UAV, and trajectory optimization for the jammer UAV.

Meng: That sounds computationally heavy. What do they say about actually making it run on a real system?

Tom: They tackle that non-convexity in some of those subproblems using successive convex approximation technology to get an estimate of the achievable rates, specifically approximating the sensing rate and the communication rate using their first order Taylor expansion.

Jane: That means they are approximating the actual performance curves to find a good solution without having to solve every single piece perfectly at once.

Lu: Plus, in that beamforming subproblem for the jammer UAV, they relax a rank one constraint into a semi definite relaxation optimization problem so it becomes solvable using standard techniques.

Meng: Okay, so what are the results? Are these approximations accurate enough for real deployment?

Tom: The computational complexity of their proposed algorithm is summarized as O lmax (N2 t N) three point five + ((K + M)N) three point five log(one/epsilon), which shows it scales reasonably well with the number of time slots N, transmit power budget, and the desired error level epsilon.

Jane: Numerically, they show that their proposed scheme outperforms other benchmarks when measuring the average secrecy rate across various scenarios and parameters.

Lu: The performance isn't totally robust either; they found it is sensitive to factors like Nt, Delta, and P fly ave.

Meng: So if I were trying to deploy this on a real drone platform, what’s the biggest practical limitation they flagged?

Tom: They point out that their model assumes a line-of-sight channel for both communication and sensing, which might not hold up in some environments.

Title and authors: Jane: And they also mention that the model focuses on a single eavesdropper, which is a simplification compared to real-world scenarios where you might have multiple potential listeners.

Lu: Their paper explicitly states that their work adopts a line-of-sight two aerial ground model with only one eavesdropper.

Tom: So, what’s the big picture implication for how we think about this technology?

Jane: It shows that by jointly optimizing all these things—scheduling, time slots, power, and movement—we can get a much better secrecy rate than if you just optimized the communication part or the sensing part separately.

Lu: The core contribution here is proposing this holistic framework where the base station UAV and the jammer UAV cooperate to execute all three tasks simultaneously under realistic physical constraints.

Meng: If we were running this on a production system, what’s the immediate next step for deployment?

Tom: I think the next logical step would be testing their framework against more complex scenarios involving multiple mobile eavesdroppers under a probabilistic line-of-sight channel model with three dimensional trajectory optimization.

Jane: That sounds like a solid direction because dealing with imperfect location knowledge and multiple listeners is definitely the next big hurdle in making this practical.

Lu: And from my side, I think the decomposition strategy they used, breaking it down into those six specific subproblems, is a really good blueprint for handling these coupled non-convex problems in future research.

Meng: I just wonder about the training cost for the AI part of their algorithm; how much overhead does that iterative optimization introduce compared to a simpler scheduling algorithm?

Tom: That’s a fair question, Meng, but they show that the complexity is manageable, and it delivers better results in terms of secrecy rate maximization under these tough constraints.

Jane: So to wrap up this discussion on "Resource Allocation for Secure Dual-UAV-Assisted ISAC System," we see a framework that tackles the difficulty of simultaneous communication and sensing by jointly optimizing all the scheduling, power, and movement parameters across two drones.

Lu: It moves away from idealized assumptions by incorporating propulsion energy limits and realistic sensing thresholds directly into the optimization process.

Meng: From my end, I’m focused on what breaks at scale; they have to show us that this iterative method doesn't require an unreasonable amount of real-time computation when the number of users gets large.

Lalam: I can confirm that based on the literature, this approach is a solid way to handle the dual objective of maximizing secrecy rate while respecting those physical limitations.

Tom: Exactly, so we’ve seen how they use block coordinate descent and successive convex approximation to manage a non-convex problem in a very practical setting.

Jane: It’s really about showing that this integrated sensing and communication approach can deliver better security performance when you have these dual-UAV systems working together.

The paper's summary: Tom: So we’re looking at how this paper says they tackle that whole mess of simultaneous communication and sensing when you have two drones working together for security.

Jane: Right, so the main idea is that you have one drone doing the talking to your users and another drone doing the radar sensing to find eavesdroppers, and they have to coordinate their movements and power use perfectly.

Lu: They’re trying to maximize this secrecy rate—which is basically how much of your data stays secret from the eavesdropper—while keeping track of all these real-world physical limits like how fast the drones can fly and how much energy they have left.

Meng: So, for someone who actually has a drone system running, what does this summary mean in plain English?

Tom: It means they’ve mapped out the whole optimization problem—user scheduling, time slots, power levels for both drones, and their flight paths—and found a way to solve it iteratively without getting totally stuck in that tangled non-convex mess.

Jane: They use this block coordinate descent method to break the giant problem down into six smaller pieces they can handle one by one.

Lu: And for those tricky parts where things aren't smooth, like when you’re approximating the achievable rates, they use successive convex approximation to get a good estimate using just the first order Taylor expansion of those rate equations.

Tom: That makes sense because solving the whole thing at once would be impossible given all those constraints and variables.

Jane: It shows that by making these specific choices for how much time is spent sensing versus communicating, and where each drone flies, you can get a significantly better secrecy performance than if you just optimized communication or sensing separately.

Lu: The core of this paper is the joint optimization—making sure the user scheduling fits with the time allocation, which fits with the power they use to talk and sense, all while respecting those velocity and energy limits.

Meng: What I’m seeing here is that it’s not just theoretical; they are trying to build a practical framework for how these two aerial vehicles actually cooperate in real-time under tight operational constraints.

Tom: That’s the point; it moves beyond just saying "we can do this" and shows the actual mechanics of how to achieve a high average secrecy rate with this dual-UAV setup.

Jane: So, if you were listening and only cared about one thing, like maximizing security, this paper tells you that coordinating all those moving parts—the scheduling and the flight paths—is exactly what’s needed for success.

Lu: And while they do mention limitations, they flag that right now their model assumes a line of sight channel for both communication and sensing with just one eavesdropper.

Meng: So, the practical implication is that this framework is a strong starting point, but you have to be ready to adjust it when you introduce multiple eavesdroppers or if the channel becomes more complex than line of sight.

The paper's improvements: Tom: Okay, so we’re looking at what the authors suggest they should do next to make this system even better than what they just described on page three.

Jane: Right, so they aren't just stopping there with their current model; they are pointing out that their setup is a bit too simple for real life.

Lu: They’re suggesting moving toward a more robust framework where the AI learns and generates distributions of worst-case eavesdropper locations within an uncertain region.

Meng: What does that mean for deployment? Does it mean we need to spend way more time training the AI to predict bad locations?

Tom: It means they’re tackling that imperfect location knowledge problem head-on; instead of assuming we know where the eavesdropper is, the AI learns how likely those locations are within a certain area.

Jane: That addresses one of those major challenges we talked about earlier—the fact that in reality, you never know exactly where the listener is.

Lu: They also suggest using this AI to jointly design trajectories and beamforming to optimize secrecy performance under conditions that are actually realistic.

Meng: So, it’s not just a static plan for movement; the system needs to be actively adapting its path and its signal direction based on what the AI predicts about the environment.

Tom: That takes the system from a fixed optimization problem to something more dynamic, which is necessary when things are moving fast in real-world drone operations.

Jane: And they also propose enabling the reliable transmission of that sensed eavesdropper information from your base station drone to your friendly jammer drone.

Lu: The jammer then broadcasts artificial noise to suppress that eavesdropper, ensuring effective jamming even when the communication itself is being delivered secretly.

Tom: It’s a two-way security mechanism; sensing for location and jamming for suppression, all coordinated by the AI.

Jane: This makes the system much more resilient because it covers both sides of the attack—knowing where they are and actively making them quiet.

Meng: That sounds like a very complex thing to implement in hardware; what’s the cost implication of adding this predictive sensing layer?

Lu: The idea is that while it increases computational demand, the benefit comes from optimizing secrecy under more realistic conditions, which is a necessary trade-off for high-security applications.

Tom: So, they are trading some raw computation for significantly better security performance when dealing with uncertainty.

Jane: And by making these two drones work together like that—one sensing and the other jamming—it shows how integrated sensing and communication can really be used to secure data in complex aerial scenarios.

Conclusion: Tom: So we’ve covered how they used block coordinate descent and successive convex approximation to solve that non-convex problem in the Resource Allocation for Secure Dual-UAV-Assisted ISAC System paper.

Jane: They proved that by jointly optimizing user scheduling, time allocation, power levels, and flight trajectories across both UAVs, you can get a better average secrecy rate under those physical constraints.

Lu: The implication is that this iterative approach is a practical way to handle these coupled variables in complex aerial systems where communication and sensing happen at the same time.

Meng: For us on the production side, it means we have a solid blueprint for an AI system that can manage those drones without getting completely stuck trying to solve every single variable perfectly all at once.

Lalam: I’m very sure this direction is sound because it directly tackles the resource constraints, like maximum velocity and energy consumption, which are real limitations in any physical deployment.

Tom: It does show that a well-structured iterative algorithm can deliver practical efficacy numerically across various scenarios tested in their experiments.

Jane: They also noted that while they used a line of sight model with one eavesdropper, the core logic for joint optimization is still very applicable to those larger problems out there.

Lu: And they explicitly pointed toward future work: exploring more complex scenarios involving multiple mobile eavesdroppers under a probabilistic line of sight channel model and three dimensional trajectory optimization.

Meng: That’s where the real engineering challenge will be; scaling that up to handle multiple moving targets while maintaining that level of accuracy.

Tom: So, to wrap this up, the Resource Allocation for Secure Dual-UAV-Assisted ISAC System paper gives us a concrete way to optimize scheduling and movement in these dual drone security setups.

Jane: It’s a strong piece of work because it shows how integrated sensing and communication can be used effectively when you have these two aerial vehicles working in tandem.

More episodes

← Home