Pith. sign in

REVIEW 4 major objections 5 minor 46 references

Covert Communications in MEC-Based Networked ISAC Systems Towards Low-Altitude Economy

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims that covert task offloading, target sensing, and UAV trajectory design can be jointly optimized in a networked ISAC-MEC system to minimize total energy while keeping wardens unable to detect transmissions.

desk verdict Competent integration, but the covertness constraint is invalid for negative μ, so the central guarantee fails as stated. read the letter →

arxiv 2507.18194 v1 pith:GUEPHI5H submitted 2025-07-24 cs.IT eess.SPmath.IT

Weihao Mao , Yang Lu , Bo Ai , Tony Q. S. Quek This is my paper
classification cs.ITeess.SPmath.IT
keywords low-altitudeeconomynetworkedISACmobileedgecomputingcovertcommunicationsUAVtrajectoryoptimizationalternatingdetectionerrorprobabilityenergyminimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper explores whether a low-altitude economy network can simultaneously receive computation tasks from UAVs, locate an unauthorized target with radar-style sensing, and keep those task-offloading transmissions hidden from wardens, all without wasting energy. The authors model a mobile edge computing server coordinating several access points whose dual-functional waveforms act as both sensing signals and jamming noise. They derive a closed-form expression for the warden's detection error probability and convert covertness into a tractable constraint. The central claim is that their alternating-optimization algorithm, combining successive convex approximation for resources and a trust-region method for UAV trajectories, minimizes total energy consumption and outperforms simpler fixed designs. A sympathetic reader would take this as a resource-allocation blueprint for covert, sensing-capable edge computing with aerial users.

What carries the argument

The load-bearing object is the warden's signal-power ratio $\mu_{l}[n]$, defined as the extra power a warden would receive if offloading were happening, divided by the power it receives during the sensing-only phase, with the access points' own waveform counted as jamming. Because the minimum detection error probability is a strictly increasing function of this ratio, the covertness requirement $\xi_{l}^{*}[n] \ge \xi_{\min}$ becomes the linearized constraint $\mu_{l}[n] \le \mu_{\max}$. That reduction is what lets communication beamformers, sensing waveforms, time splits, computation allocations, and UAV trajectories be optimized inside convex-approximation loops.

What would settle it

Directly evaluate the claimed monotonicity by computing the minimum detection error probability in equation (21) over the feasible range of $\mu_l[n]$: if the detection error probability is not strictly increasing in $\mu_l[n]$, the transformation of the covertness constraint into $\mu_l[n] \le \mu_{\max}$ fails. A second check is to run the proposed algorithm with warden channels perturbed by, say, ten percent error and observe whether the achieved detection error probability still meets $\xi_{\min}$.

Watch

Extended reading notes

Core claim

The core discovery is that covertness, sensing, and computation can be coupled in one optimization problem without making it unsolvable. Each time slot is split into an offloading-sensing phase and a computing-sensing phase, and the access points' joint waveform simultaneously illuminates the target for localization and jams the wardens, so the same transmitted energy pays for both sensing and covertness. The paper shows that the minimum detection error probability at each warden is a monotone function of a single ratio, the covert-plus-jamming power to the noise-plus-sensing power, which turns the covertness requirement into a linear constraint. On that basis, the total energy minimization over beamformers, sensing waveforms, computation resources, time splits, and UAV trajectories is decomposed into parallel per-slot resource-allocation problems solved by successive convex approximation and a trajectory subproblem solved by trust-region updates. Simulations show optimized trajectories bending toward access points and away from wardens, and the proposed design beating straight-flight, fixed-power, fixed-time, and full-offloading benchmarks.

Load-bearing premise

The design assumes the network knows the wardens' channel gains and noise powers exactly, because those values feed directly into the detection-error constraints; if that knowledge is wrong, the promised covertness guarantee cannot actually be enforced.

Editorial extensions

If this is right

  • If the algorithm's near-optimality is accepted, an operator can jointly choose offloading powers, sensing beams, computation rates, phase durations, and UAV routes under a single energy objective while keeping every warden's detection error probability above threshold.
  • Because the access point waveform is dual-functional, raising the radar SINR requirement forces more sensing power and sensing time, which shrinks the offloading phase and pushes tasks onto less efficient on-board UAV computation; the paper's simulations show a sharp energy increase as the radar SINR threshold passes 0.5.
  • The optimized UAV trajectories curve toward access points and away from wardens, which improves offloading link quality and covertness at the cost of extra propulsion energy and less favorable offloading geometry.
  • The offloading ratio depends strongly on the MEC server's computation cost coefficient; when that coefficient rises above roughly one tenth of the UAV's coefficient, the simulations show edge computing becomes markedly less cost-effective.
  • The proposed design's advantage over fixed time assignment grows as the radar SINR requirement tightens, because flexible phase splitting adds a degree of freedom that static schedules lack.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper, the same monotone-ratio reduction of detection error probability may extend to wardens with multiple antennas or uncertain noise power, where the covertness constraint would become a robust or quasi-convex condition that successive convex approximation could still handle.
  • The strongest exposed premise is perfect knowledge of warden channels and noise powers; if those are uncertain, the detection error guarantee would need a worst-case reformulation, which would likely raise the reported energy consumption.
  • The sharp energy transition at radar SINR thresholds around 0.5 suggests a practical scheduler could deliberately operate just below such thresholds, accepting a small sensing margin to avoid disproportionately large energy costs.
  • A testable extension would replace the fixed target sample set with an online target tracker and check whether the optimized beams and trajectories remain energy-efficient when the target location changes within each slot.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper studies a multi-UAV MEC-based networked ISAC system in which multiple access points simultaneously receive computation tasks from UAVs, perform target sensing, and emit dual-functional jamming/sensing waveforms to hide the UAV offloading from multiple wardens. A total energy minimization problem is formulated over transmit beamformers, sensing waveforms, computation resources, time splitting, and UAV trajectories, subject to covertness (DEP), radar SINR, MEC QoS, and collision constraints. The authors propose an alternating-optimization framework: a successive convex approximation method for joint resource allocation and a trust-region trajectory optimization method. Simulations compare the proposed design with power-allocation, straight-flight, fixed-time-assignment, and full-offloading benchmarks.

Significance. If correct, the paper would offer a useful integration of covert communications, ISAC, MEC, and trajectory optimization for low-altitude economy systems, with a plausible energy-efficiency advantage over the presented benchmarks. The problem formulation is comprehensive and the simulation study is extensive. However, the correctness of the central covertness guarantee is undermined by a flaw in the DEP reformulation, so the main claim that the proposed design maintains covertness while minimizing energy is not currently established.

major comments (4)
  1. [Section II-C, Proposition 1 and Eqs. (19)-(24)] The DEP reformulation is only valid for μ_l[n] > 0, i.e., λ1_l[n] > λ0_l[n]. For μ_l[n] ∈ (-1, 0) the likelihood ratio in (16) is decreasing in |y_W,l[n]|^2, so the optimal test is the reverse inequality; equation (21) then yields values exceeding 1, which is impossible, and the true minimum DEP is 1 - (1+μ)^{-(1+μ)/μ} + (1+μ)^{-1/μ}, which can be far below ξ_min. The monotonicity claim in (23) is false on (-1,0) because ln(1+μ)/μ^2 is negative there. Since no constraint in P0, P1, or P2 enforces μ_l[n] ≥ 0, the feasible set contains points with negative μ, and constraint (24) is not equivalent to the intended DEP guarantee. This is a load-bearing error: the proposed algorithm may output designs that violate the covertness requirement. The authors must either prove Proposition 1 under an explicit assumption μ_l[n] ≥ 0 and add that constraint, or correctly handle the μ<0 regime, and then re-run the design and simulations under the corrected constraint.
  2. [Section III-A, reformulation P1 to P1.1] The equivalence of P1 and P1.1 is asserted but not proved. The exponential substitutions and the auxiliary variables a0, a1, b, r_k, γ_k, ζ_k, p0, p1, p2 introduce constraints whose tightness at optimality is not established. For instance, constraints (32)-(33) and (41)-(42) can be satisfied with slack, and it is not demonstrated that the optimal value of P1.1 equals that of P1. A short proof that, at a minimizer of P1.1, one can always choose the auxiliary variables such that the original constraints hold with equality (or at least that any feasible P1.1 solution maps to a feasible P1 solution with no larger objective) is needed to support the claim that Problem P1.2 solves P1.
  3. [Section III-B, Algorithm 2] The convergence of the trust-region algorithm is only justified by citing [46] and stating that for sufficiently small ω the convergence is guaranteed. The paper does not prove that the sequence of objective values of Problem P2.2, with the imposed trust region (76), produces a monotone decrease of the original objective (55a), nor that the first-order Taylor approximations in (71)-(72) remain valid over the trust region. Since Algorithm 2 is a central component of the proposed solution, a formal convergence argument or a detailed reference to a theorem that exactly applies to this setting is required.
  4. [Section II-C, Proposition 1 proof status] Proposition 1 is a central result of the paper, but its proof is omitted and deferred to 'Appendix A in [36]'. The statement is not a trivial modification of [36] because the expressions for λ0_l[n] and λ1_l[n] differ and the domain of μ_l[n] is not discussed. The paper must include a complete proof (or a detailed proof sketch) that explicitly states the assumptions under which (19)-(21) hold, in particular the sign of μ_l[n] and the corresponding optimal threshold test.
minor comments (5)
  1. [Abstract] The abstract contains the typo 'conver t transmission'; it should read 'covert transmission'.
  2. [Section II-C, Eq. (20)] The domain of μ_l[n] is not stated; since λ0_l[n] > 0, one has μ_l[n] > -1, but the paper never mentions this restriction, which is relevant to the DEP discussion.
  3. [Fig. 2] The figure uses the symbol 'ND' without defining it; the text uses N for the number of time slots. Please align the notation.
  4. [Section IV, Fig. 4 discussion] The statement that the power allocation design is infeasible for PAP,max = 30 W is not explained; the reader is left to infer why increasing the AP budget to 90 W is needed.
  5. [Conclusion] The word 'reveled' in the final paragraph should be 'revealed'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the DEP constraint is imported from external prior work [36] and the optimization chain is not reduced to its own inputs.

full rationale

The paper's central covertness step is Proposition 1, which gives the warden's optimal threshold and minimum DEP. The proof is explicitly delegated to Appendix A of [36], an external paper by different authors, with only the expressions for lambda_0 and lambda_1 modified to match the present system model. This is an application of an independently published result, not a self-citation chain or a construction that assumes the target conclusion. The subsequent transformation of the DEP constraint into (24) and (39) is a direct algebraic rewrite of that external expression, not a fitted parameter later relabeled as a prediction. The optimization objective and constraints are assembled from standard channel, propulsion, computation, radar SINR, and covertness models, with no data-fitting step. Self-citations [5], [12], and [37] appear only for motivational context and the discrete path-planning modeling convention; none of them carries the load of the claimed energy savings or the covertness guarantee. The SCA and trust-region machinery is anchored to external references [44], [46]. The skeptical concern that the monotonicity claim in (23) fails for negative mu_l[n] is a potential mathematical correctness issue, not a circularity, because the alleged reduction would fail on independent mathematical grounds rather than by construction. The perfect-CSI assumption and the future-work note about imperfect CSI in Section V are limitations, but they do not make the derivation circular. Overall, the paper's derivation is self-contained relative to its external DEP source and does not exhibit self-definitional, fitted-input, or load-bearing self-citation patterns.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard LoS channel models, perfect channel state information including warden links, and the energy-detection DEP expression borrowed from [36]. No free parameters are fitted to data; all simulation constants are scenario inputs from prior literature. The main unstated load-bearing premises are perfect CSI and exact knowledge of warden channels and noise powers, acknowledged only as future work on imperfect CSI.

assumptions (6)
  • domain assumption All four link types (offloading, warden, jamming, cascaded sensing) are modeled as deterministic line-of-sight channels.
    Section II-A states 'these links are assumed to be LoS as they represent either air-ground or air-air transmission.' This removes fading and multipath and makes the channel matrices exact functions of positions.
  • domain assumption The system has perfect knowledge of all channels, including warden channels and warden noise power.
    The covert constraints (24) and (39) use exact h_{l,k}[n], g_l, and sigma_l^2; no uncertainty or estimation error is modeled. Section V lists imperfect CSI only as future work.
  • domain assumption The warden uses an energy-detection likelihood ratio test with optimal threshold, and the DEP expression (21) from [36] applies to the modified lambda0 and lambda1.
    Section II-C invokes the Neyman-Pearson criterion and Proposition 1, whose proof is deferred to Appendix A of [36] with 'modifications only to the expressions for lambda0 and lambda1.'
  • domain assumption UAV positions are constant within each time slot, with N large enough that Delta T is small.
    Section II: 'N is assumed to be large enough such that Delta T is sufficiently small and the UAV locations can be assumed to be unchanged during each time slot.'
  • standard math The optimization uses convex approximations based on first-order Taylor expansions and trust-region convergence from [46].
    Section III uses SCA and trust-region steps, with convergence cited to [46]; no proof of convergence to a stationary point is given in this paper.
  • standard math Exponential tail probabilities for complex Gaussian noise underlie the DEP calculation.
    The likelihood ratio test in Section II-C assumes |y_W|^2 is exponentially distributed under each hypothesis, which follows from the Gaussian noise model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Covert Communications in MEC-Based Networked ISAC Systems Towards Low-Altitude Economy." pith.science (2026). https://pith.science/paper/GUEPHI5H

@misc{pith2026250718194,
  author       = {Pith},
  title        = {Pith review of: Covert Communications in MEC-Based Networked ISAC Systems Towards Low-Altitude Economy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GUEPHI5H}},
  note         = {Machine review of arXiv:2507.18194}
}
read the original abstract

Low-altitude economy (LAE) is an emerging business model, which heavily relies on integrated sensing and communications (ISAC), mobile edge computing (MEC), and covert communications. This paper investigates the convert transmission design in MEC-based networked ISAC systems towards LAE, where an MEC server coordinates multiple access points to simultaneously receive computation tasks from multiple unmanned aerial vehicles (UAVs), locate a target in a sensing area, and maintain UAVs' covert transmission against multiple wardens. We first derive closed-form expressions for the detection error probability (DEP) at wardens. Then, we formulate a total energy consumption minimization problem by optimizing communication, sensing, and computation resources as well as UAV trajectories, subject to the requirements on quality of MEC services, DEP, and radar signal-to-interference-and-noise ratio, and the causality of UAV trajectories. An alternating optimization based algorithm is proposed to handle the considered problem, which decomposes it into two subproblems: joint optimization of communication, sensing, and computation resources, and UAV trajectory optimization. The former is addressed by a successive convex approximation based algorithm, while the latter is solved via a trust-region based algorithm. Simulations validate the effectiveness of the proposed algorithm compared with various benchmarks, and reveal the trade-offs among communication, sensing, and computation in LAE systems.

Figures

Figures reproduced from arXiv: 2507.18194 by the authors.

Figure 1
Figure 1. Illustration of an MEC-based networked ISAC system. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. illustrates the time slot assignment at UAVs and APs. Throughout each time slot, UAVs conduct local computation. However, ISAC APs divide their time slots into two phases: 1) offloading-sensing phase, where APs receive the computation tasks offloaded by UAVs while performing target localization, and 2) computing-sensing phase, where the MEC server ex￾ecutes the offloaded computation tasks while APs continue target l… view at source ↗
Figure 3
Figure 3. Trajectories of UAVs under various warden locations [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Comparison among the proposed, power allocation, an [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Comparison between the proposed and fixed time assign [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 7
Figure 7. Figure 7: The total energy consumption and offloading ratio ver [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

46 extracted references · 45 canonical work pages

  1. [36]

    Joint beamforming and UA V trajectory optimizat ion for covert communications in ISAC networks,

    D. Deng, W. Zhou, X. Li, D. B. da Costa, D. W. K. Ng, and A. Na l- lanathan, “Joint beamforming and UA V trajectory optimizat ion for covert communications in ISAC networks,” IEEE Trans. Wireless Commun., vol. 24, no. 2, pp. 1016-1030, Feb. 2025

    work page 2025

  2. [46]

    A. R. Conn, N. I. M. Gould, and P . L. Toint, Trust-Region Methods . Philadelphia, PA, USA: SIAM, 2000

    work page 2000

  3. [1]

    Trustworthy intelligent networks for lo w-altitude econ- omy,

    M. Song et al., “Trustworthy intelligent networks for lo w-altitude econ- omy,” IEEE Commun. Mag. , vol. 63, no. 7, pp. 72-79, July 2025

    work page 2025

  4. [2]

    The potential of low-alt itude airspace: The future of urban air transportation,

    H. Huang, J. Su, and F. -Y . Wang, “The potential of low-alt itude airspace: The future of urban air transportation,” IEEE Trans. Intell. V eh. , vol. 9, no. 8, pp. 5250-5254, Aug. 2024

    work page 2024

  5. [3]

    Efficient DOA esti mation method for reconfigurable intelligent surfaces aided UA V sw arm,

    P . Chen, Z. Chen, B. Zheng, and X. Wang, “Efficient DOA esti mation method for reconfigurable intelligent surfaces aided UA V sw arm,” IEEE Trans. Signal Process. , vol. 70, pp. 743-755, 2022

    work page 2022

  6. [4]

    Integrated sensing and communication f or low altitude economy: Opportunities and challenges,

    Y . Jiang et al., “Integrated sensing and communication f or low altitude economy: Opportunities and challenges,” early accessed in IEEE Com- mun. Mag

    work page

  7. [5]

    UA V -assisted communicat ions in SAGIN-ISAC: Mobile user tracking and robust beamforming,

    W. Mao, Y . Lu, G. Pan, and B. Ai, “UA V -assisted communicat ions in SAGIN-ISAC: Mobile user tracking and robust beamforming,” IEEE J. Sel. Areas Commun. , vol. 43, no. 1, pp. 186-200, Jan. 2025

    work page 2025

  8. [6]

    Ubiquitous UA V communic ation enabled low-altitude economy: Applications, techniques, and 3GPP’s efforts,

    D. He, W. Y uan, J. Wu, and R. Liu, “Ubiquitous UA V communic ation enabled low-altitude economy: Applications, techniques, and 3GPP’s efforts,” early accessed in IEEE Netw

    work page

Show all 46 references
  1. [7]

    Integrated se nsing and communications for low-altitude economy: A deep reinforce ment learning approach,

    X. Y e, Y . Mao, X. Y u, S. Sun, L. Fu, and J. Xu, “Integrated se nsing and communications for low-altitude economy: A deep reinforce ment learning approach,” early accessed in IEEE Trans. Wireless Commun

  2. [8]

    Energy-efficient STAR-RIS enhanced UA V -enabled MEC netwo rks with bi-directional task offloading,

    H. Xiao, X. Hu, W. Zhang, W. Wang, K. -K. Wong, and K. Y ang, “Energy-efficient STAR-RIS enhanced UA V -enabled MEC netwo rks with bi-directional task offloading,” IEEE Trans. Wireless Commun. , vol. 24, no. 4, pp. 3258-3272, Apr. 2025. 13 ∂aU(φ l,k[n]) ∂x U k [n] = [ 0, . . . ,...

    2025

  3. [9]

    Joint power alloca tion and trajectory design for UA V -enabled covert communication,

    P . Wu, X. Y uan, Y . Hu, and A. Schmeink, “Joint power alloca tion and trajectory design for UA V -enabled covert communication,” IEEE Trans. Wireless Commun., vol. 23, no. 1, pp. 683-698, Jan. 2024

    2024

  4. [10]

    Uncovering the iceberg in the sea: Fundam entals of pulse shaping and modulation design for random ISAC signals,

    F. Liu et al., “Uncovering the iceberg in the sea: Fundam entals of pulse shaping and modulation design for random ISAC signals,” IEEE Trans. Signal Process. , vol. 73, pp. 2511-2526, 2025

    2025

  5. [11]

    STAR-RIS enabled ISAC systems with RSMA: Joint rate splitting and beamforming optimization,

    Y . Liu et al., “STAR-RIS enabled ISAC systems with RSMA: Joint rate splitting and beamforming optimization,” early accessed in IEEE Trans. Cogn. Commun. Netw

  6. [12]

    Communication-sensing region for cell-free massive MIMO ISAC sys- tems,

    W. Mao, Y . Lu, C. -Y . Chi, B. Ai, Z. Zhong, and Z. Ding, “Communication-sensing region for cell-free massive MIMO ISAC sys- tems,” IEEE Trans. Wireless Commun. , vol. 23, no. 9, pp. 12396-12411, Sept. 2024

    2024

  7. [13]

    Reshaping the ISAC tradeoff under OFDM signaling: A probabilistic con stellation shaping approach,

    Z. Du, F. Liu, Y . Xiong, T. X. Han, Y . C. Eldar, and S. Jin, “ Reshaping the ISAC tradeoff under OFDM signaling: A probabilistic con stellation shaping approach,” IEEE Trans. Signal Process. , vol. 72, pp. 4782-4797, 2024

    2024

  8. [14]

    Through put maximization for UA V -enabled integrated periodic sensing and commu- nication,

    K. Meng, Q. Wu, S. Ma, W. Chen, K. Wang, and J. Li, “Through put maximization for UA V -enabled integrated periodic sensing and commu- nication,” IEEE Trans. Wireless Commun. , vol. 22, no. 1, pp. 671-687, Jan. 2023

    2023

  9. [15]

    Beamforming design and tr ajectory optimization for UA V -empowered adaptable integrated sensing and com- munication,

    C. Deng, X. Fang, and X. Wang, “Beamforming design and tr ajectory optimization for UA V -empowered adaptable integrated sensing and com- munication,” IEEE Trans. Wireless Commun. , vol. 22, no. 11, pp. 8512- 8526, Nov. 2023

    2023

  10. [16]

    Hybrid-RIS empowered UA V -assisted ISA C systems: Transfer learning-based DRL,

    P . Saikia, A. Jee, K. Singh, W. -J. Huang, A. -A. A. Boulog eorgos, and T. A. Tsiftsis, “Hybrid-RIS empowered UA V -assisted ISA C systems: Transfer learning-based DRL,” early accessed in IEEE Trans. Commun

  11. [17]

    Cooperative ISAC-empowered low-altit ude economy,

    J. Tang et al., “Cooperative ISAC-empowered low-altit ude economy,” IEEE Trans. Wireless Commun., vol. 24, no. 5, pp. 3837-3853, May 2025

    2025

  12. [18]

    Netwo rked ISAC based UA V tracking and handover towards low-altitude econo my,

    Y . Feng, C. Zhao, H. Luo, F. Gao, F. Liu, and S. Jin, “Netwo rked ISAC based UA V tracking and handover towards low-altitude econo my,” early accessed in IEEE Trans. Wireless Commun

  13. [19]

    Networked ISAC for l ow-altitude economy: Coordinated transmit beamforming and UA V trajectory design,

    G. Cheng, X. Song, Z. Lyu, and J. Xu, “Networked ISAC for l ow-altitude economy: Coordinated transmit beamforming and UA V trajectory design,” early accessed in IEEE Trans. Commun

  14. [20]

    Joint traj ectory optimization and resource allocation in UA V -MEC systems: A Lyapunov- assisted DRL approach,

    Y . Chen, Y . Y ang, Y . Wu, J. Huang, and L. Zhao, “Joint traj ectory optimization and resource allocation in UA V -MEC systems: A Lyapunov- assisted DRL approach,” IEEE Trans. Services Comput. , vol. 18, no. 2, pp. 854-867, Mar.-Apr. 2025

    2025

  15. [21]

    Latency- aware task scheduling in software-defined edge and cloud com puting with erasure-coded storage systems,

    J. Tang, M. M. Jalalzai, C. Feng, Z. Xiong, and Y . Zhang, “ Latency- aware task scheduling in software-defined edge and cloud com puting with erasure-coded storage systems,” IEEE Trans. Cloud Comput. , vol. 11, no. 2, pp. 1575-1590, 1 Apr.-June 2023

    2023

  16. [22]

    Ling et al., ”Exploring MEC server strategy in blockc hain networks: Mining for mobile users or for self,” IEEE Trans

    X. Ling et al., ”Exploring MEC server strategy in blockc hain networks: Mining for mobile users or for self,” IEEE Trans. Mobile Comput. , vol. 24, no. 7, pp. 6030-6044, July 2025

    2025

  17. [23]

    A PDDQNLP algorithm for energy efficient computation offloadi ng in UA V - assisted MEC,

    N. Lin, H. Tang, L. Zhao, S. Wan, A. Hawbani, and M. Guizan i, “A PDDQNLP algorithm for energy efficient computation offloadi ng in UA V - assisted MEC,” IEEE Trans. Wireless Commun., vol. 22, no. 12, pp. 8876- 8890, Dec. 2023

    2023

  18. [24]

    Latenc y minimization for UA V -enabled URLLC-based mobile edge comp uting systems,

    Q. Wu, M. Cui, G. Zhang, F. Wang, Q. Wu, and X. Chu, “Latenc y minimization for UA V -enabled URLLC-based mobile edge comp uting systems,” IEEE Trans. Wireless Commun. , vol. 23, no. 4, pp. 3298-3311, Apr. 2024

    2024

  19. [25]

    Multi- agent deep reinforcement learning based UA V trajectory opt imization for differentiated services,

    Z. Ning, Y . Y ang, X. Wang, Q. Song, L. Guo, and A. Jamalipo ur, “Multi- agent deep reinforcement learning based UA V trajectory opt imization for differentiated services,” IEEE Trans. Mobile Comput. , vol. 23, no. 5, pp. 5818-5834, May 2024

    2024

  20. [26]

    Resource and trajectory optimization for UA V -relay-assisted secure maritime MEC,

    F. Lu et al., “Resource and trajectory optimization for UA V -relay-assisted secure maritime MEC,” IEEE Trans. Commun. , vol. 72, no. 3, pp. 1641- 1652, Mar. 2024

    2024

  21. [27]

    Air–ground coo rdinated MEC: Joint task, time allocation and trajectory design,

    L. Wang, Y . Li, Y . Chen, T. Li, and Z. Yin, “Air–ground coo rdinated MEC: Joint task, time allocation and trajectory design,” IEEE Trans. V eh. Technol., vol. 74, no. 3, pp. 4728-4743, Mar. 2025

    2025

  22. [28]

    Join t task offloading, resource allocation, and trajectory desig n for Multi-UA V cooperative edge computing with task priority,

    H. Hao, C. Xu, W. Zhang, S. Y ang, and G. -M. Muntean, “Join t task offloading, resource allocation, and trajectory desig n for Multi-UA V cooperative edge computing with task priority,” IEEE Trans. Mobile Comput., vol. 23, no. 9, pp. 8649-8663, Sept. 2024

    2024

  23. [29]

    RIS empowered near- field covert communications,

    J. Liu, G. Y ang, Y . Liu, and X. Zhou, “RIS empowered near- field covert communications,” IEEE Trans. Wireless Commun. , vol. 23, no. 10, pp. 15477-15492, Oct. 2024

    2024

  24. [30]

    Covert prompt transmission for secure large language model services

    R. Zhang et al., “Covert prompt transmission for secure large language model services”, arXiv preprint arXiv:2504.21311 , 2025. [Online]

    2025 arXiv

  25. [31]

    UA V -enabl ed communication strategy against detection in covert commun ication with asymmetric information,

    W. Tian, J. Du, X. Ji, M. Du, G. Liu, and Z. Han, “UA V -enabl ed communication strategy against detection in covert commun ication with asymmetric information,” IEEE Trans. Wireless Commun. , vol. 23, no. 7, pp. 6921-6937, July 2024

    2024

  26. [32]

    Jammer-ai ded covert collaborative UA V communications against directio nal beam,

    W. Chen, H. Ding, S. Wang, J. Luo, and F. Gong, “Jammer-ai ded covert collaborative UA V communications against directio nal beam,” early accessed in IEEE Trans. Commun

  27. [33]

    On joi nt optimization of trajectory and phase shift for IRS-UA V assi sted covert communication systems,

    Y . Qian, C. Y ang, Z. Mei, X. Zhou, L. Shi, and J. Li, “On joi nt optimization of trajectory and phase shift for IRS-UA V assi sted covert communication systems,” IEEE Trans. V eh. Technol., vol. 72, no. 10, pp. 12873-12883, Oct. 2023

    2023

  28. [34]

    Task offloading strategy for UA V -assisted m obile edge computing with covert transmission

    Z. Hu et al. “Task offloading strategy for UA V -assisted m obile edge computing with covert transmission” Electronics 14, no. 3, Jan. 2025

    2025

  29. [35]

    A UA V -enabled integrated sen sing and covert communication system,

    Y . Zhou, P . Fan, and Z. Ma, “A UA V -enabled integrated sen sing and covert communication system,” early accessed in IEEE Trans. V eh. Technol

  30. [37]

    Energy consu mption minimization in secure multi-antenna UA V -assisted MEC net works with channel uncertainty,

    W. Mao, K. Xiong, Y . Lu, P . Fan, and Z. Ding, “Energy consu mption minimization in secure multi-antenna UA V -assisted MEC net works with channel uncertainty,” IEEE Trans. Wireless Commun., vol. 22, no. 11, pp. 7185-7200, Nov. 2023

    2023

  31. [38]

    G¨ uven and G

    E. G¨ uven and G. K. Kurt, ”Learning-to-learn the wave an gle estimation,” IEEE Trans. Commun. , vol. 72, no. 8, pp. 4961-4977, Aug. 2024

    2024

  32. [39]

    Air-to-ground communication s beyond 5G: UA V swarm formation control and tracking,

    X. Fan, P . Wu, and M. Xia, “Air-to-ground communication s beyond 5G: UA V swarm formation control and tracking,” IEEE Trans. Wireless Commun., vol. 23, no. 7, pp. 8029-8043, July 2024

    2024

  33. [40]

    3D deployment and e nergy efficiency optimization based on DRL for RIS-assisted air-t o-ground communications networks,

    Y . Y ao, K. Lv, S. Huang, and W. Xiang, “3D deployment and e nergy efficiency optimization based on DRL for RIS-assisted air-t o-ground communications networks,” IEEE Trans. V eh. Technol., vol. 73, no. 10, pp. 14988-15003, Oct. 2024

    2024

  34. [41]

    Computation-awa re offload- ing for DNN inference tasks in semantic communication assis ted MEC systems,

    G. Zheng, M. Wen, Z. Ning, and Z. Ding, “Computation-awa re offload- ing for DNN inference tasks in semantic communication assis ted MEC systems,” IEEE Trans. Wireless Commun. , vol. 24, no. 4, pp. 2693-2706, Apr. 2025

    2025

  35. [42]

    IRS empowered MEC system with computation offloading, reflecting design, and beamforming optimization,

    Z. Wu et al., “IRS empowered MEC system with computation offloading, reflecting design, and beamforming optimization,” IEEE Trans. Commun., vol. 72, no. 5, pp. 3051-3063, May 2024

    2024

  36. [43]

    Covert commu nication with time uncertainty in time-critical wireless networks,

    X. Lu, S. Y an,W. Y ang, M. Li, and D.W. K. Ng, “Covert commu nication with time uncertainty in time-critical wireless networks, ” IEEE Trans. Wireless Commun, vol. 22, no. 2, pp. 1116–1129, Feb. 2023

    2023

  37. [44]

    Secure UA V -enabled communication using Han–Kobayashi si gnaling,

    Z. Sheng, H. D. Tuan, A. A. Nasir, T. Q. Duong, and H. V . Poo r, “Secure UA V -enabled communication using Han–Kobayashi si gnaling,” IEEE Trans. Wireless Commun., vol. 19, no. 5, pp. 2905-2919, May 2020

    2020

  38. [45]

    Jo int placement and beamforming design in UA V enabled multi-stag e ISAC system,

    L. Xu, Q. Zhu, W. Xia, Z. Wang, T. Q. S. Quek, and H. Zhu, “Jo int placement and beamforming design in UA V enabled multi-stag e ISAC system,” early accessed in IEEE Trans. Commun

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.