Digital Twin-Guided Energy Management over Real-Time Pub/Sub Protocol in 6G Smart Cities

Berk Canberk; Kerem Gursu; Kubra Duran; Lal Verda Cakir; Sana Ullah Jan

REVIEW 4 major objections 6 minor 21 references

Digital Twin-Guided Energy Management over Real-Time Pub/Sub Protocol in 6G Smart Cities

T0 review · 4 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A digital-twin overlay with RL-scheduled data updates cuts smart-city IoT latency by 37% and energy by 30%.

desk verdict Plausible DT/RTPS/DDPG integration with real architectural content, but the 37%/30% numbers rest on an unspecified simulator and one reward term is identically zero—fixable, but not currently auditable. read the letter →

arxiv 2508.18516 v1 pith:FB5PGRAM submitted 2025-08-25 cs.NI

Kubra Duran , Lal Verda Cakir , Sana Ullah Jan , Kerem Gursu , Berk Canberk This is my paper · ORCID

classification cs.NI

keywords digitaltwinenergyefficiencylowlatencyreinforcementlearningDDPGRTPSsmartcitiesIoT

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

The paper proposes a four-layer digital-twin architecture for 6G smart cities in which IoT devices are mirrored as DDS overlay nodes and kept in sync over the Real-Time Publish/Subscribe (RTPS) protocol. On top of this, a DDPG reinforcement-learning agent picks a continuous transmission time for every sensor, rewarded for high remaining battery, low communication delay relative to a threshold, and respecting a minimum gap between consecutive transmissions. The authors' central claim is that this combination yields 37% lower 95th-percentile latency and 30% lower energy consumption than four prior methods in their simulated one-service and three-service smart-city scenarios. A sympathetic reader would care because the result suggests that energy savings do not have to come from discrete offloading choices; continuously scheduling when devices wake up and publish can serve both goals.

What carries the argument

The carrying mechanism is the DDPG-based RL engine operating on a continuous action space where each action a_i is the minute of the day device i transmits. The composite reward R_total = R_energy + R_timeliness + R_consecutive ties the policy to the two goals: R_energy sums the battery levels resulting from chosen actions, R_timeliness penalizes delays above D_threshold, and R_consecutive penalizes transmissions closer than Δt_min. The RTPS/DDS distributed overlay is the supporting infrastructure that keeps twin states and application updates synchronized; the reward-shaped continuous scheduler is what produces the reported gains.

What would settle it

Run the same trained DDPG scheduling policy on physical IoT devices, or in a packet-level simulator with a realistic radio energy model, and compare end-to-end latency and battery drain against Partial, REWIRE, ECCO, and D2LIEO under identical traffic; if the 37% and 30% margins shrink or disappear, the simulator's action-to-delay-to-battery mapping was the cause.

Watch

Extended reading notes

Core claim

The central claim is that continuous reinforcement-learning control of data update times, rather than discrete offload-or-local decisions, is the missing mechanism for jointly satisfying low-latency and energy-efficiency targets in digital-twin-managed 6G IoT. The framework routes twin synchronization and data exchange over RTPS on a distributed DDS overlay, then lets a DDPG actor output each device's transmission time in [0,1440] minutes. The reward function couples remaining battery, a per-device delay threshold (180 ms in the simulations), and a penalty for violating the minimum inter-transmission interval; the authors report that this policy keeps 95th-percentile latency stable as commun

Load-bearing premise

The simulation's implicit mapping from a scheduled transmission time to both communication delay and battery drain is the load-bearing premise; if that mapping does not match real device behavior, the reported gains are artifacts of the simulator.

Editorial extensions

If this is right

The same scheduling principle should generalize to any digital-twin IoT scenario where data freshness and battery life trade off, not only the three smart-city services tested.
Because actions are continuous wake-up times rather than offload decisions, the approach can be layered on top of existing publish/subscribe middleware without changing device hardware.
The reported figures imply that discovery-phase and data-exchange-phase latency can both be kept near a threshold while scaling to at least 600 publishers.
The 180 ms delay threshold acts as a tunable knob: changing it should shift the operating point between energy savings and timeliness, giving operators a direct service-level control.

Reading between the lines

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

Beyond the paper, the same DT-guided reward could be adapted to optimize freshness metrics such as Age of Information directly, with D_threshold replaced by an AoI budget.
A natural testable extension would be replacing the fixed Δt_min constraint with a learned or traffic-dependent minimum interval, which could improve responsiveness in bursty smart-city conditions.
Because the comparison is against offloading-based baselines, the largest gains may come from avoiding redundant transmissions entirely; measuring packet-level overhead in a real DDS deployment would show whether RTPS discovery messages erode the reported energy margin.
Since the RL engine is decoupled from the data plane, the policy could be retrained on live city data and deployed incrementally, a path the paper leaves implicit.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: this is a legitimate integration effort, not a restatement of prior work, but the two headline numbers can't be checked as written. The new piece is the combination—DDS/RTPS overlay for digital twin state distribution, a DDPG agent choosing continuous data-update times, three smart-city services, and four baselines. The RTPS operation set in Table I is concrete, the layered architecture is coherent, and the simulation uses real datasets. The authors clearly know the problem and the domain.

The soft spots are structural. Equation (4)'s consecutive-transmission penalty is identically zero: Equation (1) already enforces a'_i >= a_i + Δt_min, so max(0, Δt_min - (a'_i - a_i)) never fires. That is not a cosmetic typo; it removes one of the three reward terms. More importantly, the mapping from action a_i to the delay D_i in Equation (3) is never defined—no traffic model, no topology, no relationship between scheduled transmission time and communication delay. Energy gets the same treatment: 'each decrease in battery level is energy consumption' but there is no battery curve or transmission energy model. Without those two functions the reward cannot be evaluated, so Figures 3 and 4 are not interpretable. The 37%/30% numbers may be real inside the simulator, but the simulator is a black box. The baselines also solve different tasks (Partial and REWIRE do topology discovery; ECCO and D2LIEO do offloading), so the comparison is not apples-to-apples. The citation pattern is fine—self-citations point to directly relevant prior work.

None of this sinks the architecture. The RTPS-overlay idea is worth keeping, and continuous RL control of update times is a sensible direction. But the paper needs a proper system model, a fixed reward equation, an ablation, and ideally code/data. The reader's conditional verdict is about right.

Who it is for: researchers working on digital-twin management in IoT/6G, especially those considering DDS/RTPS as the data plane. It deserves a serious referee—the architecture is concrete and the problem is real—but it needs major revision before the quantitative claims should be trusted.

Referee Report

4 major / 6 minor

Summary. The paper proposes a four-layer digital twin (DT) framework for energy management in 6G smart cities: a DDS/RTPS distributed overlay for twin synchronization, a management layer with a What-if engine and a DDPG-based RL engine that outputs continuous transmission times for IoT devices. The reward comprises a battery-level reward, a timeliness penalty, and a consecutive-transmission penalty. Simulations over one-service and three-service smart-city scenarios claim a 37% lower 95th-percentile latency and a 30% reduction in energy consumption compared with Partial [11], REWIRE [13], ECCO [17], and D2LIEO [18]. The central claim is that continuous RL-based update scheduling over RTPS can substantially improve both latency and energy efficiency.

Significance. The architectural contribution is relevant: the RTPS operation set, domain-based overlay creation, and the use of DDPG for continuous action generation are useful design elements, and grounding the three services in public datasets is a strength. If the quantitative claims were reproducible, the paper would demonstrate a practical benefit of RL-driven update scheduling over a DDS/RTPS overlay. However, the simulation models that support the headline numbers are not specified: the delay D_i and battery level B_i are not defined as functions of the action, and the consecutive-transmission penalty in Eq. (4) is identically zero under Eq. (1). As a result, the reported 37%/30% gains are not auditable and cannot currently be accepted as system properties.

major comments (4)

[Section III-D, Eqs. (1) and (4)] R_consecutive is identically zero. Equation (1) defines a'_i = max(a_i + Δt_min, last transmission time), so a'_i - a_i ≥ Δt_min for every action. Hence max(0, Δt_min - (a'_i - a_i)) = 0 always. The claimed 'penalty for consecutive transmissions' therefore never contributes to the total reward. Either the constraint is already enforced by Eq. (1), in which case Eq. (4) is redundant, or the penalty is not implemented as written. This directly affects the claim of a 'novel' reward function.
[Section III-D, Eq. (3)] D_i is never defined as a function of the action a_i, message size, network load, or topology. The text states only that D_i is the communication delay for sensor i and that the timeliness reward penalizes D_i > D_threshold. Without a delay model, R_timeliness cannot be evaluated, and Figures 3/4 cannot be reproduced or interpreted. The authors must specify how a_i affects D_i, including any queueing, propagation, or protocol-related delay components.
[Section IV, Eq. (2)] B_i(a_i) is also undefined. The paper says 'each decrease in the battery level is considered as the energy consumption value' but gives no battery discharge curve, no transmission energy model, and no relationship between a_i and battery drain. Without this, the total energy consumption comparison in Fig. 4 is unsupported. A concrete energy model—e.g., energy per transmission, sleep/idle/active power states, and how the scheduled time affects the number of transmissions—is required.
[Section IV] The simulation setup is underspecified. The paper does not state the number of devices per domain, network topology, traffic generation model, RTPS/DDS QoS parameters, or the implementation details of the baselines [11], [13], [17], [18] in the same simulator. These are load-bearing for the comparative claims. The authors should either release the simulation code or provide a complete, step-by-step model specification so that the reported 37% and 30% gains can be independently checked.

minor comments (6)

[Section III-D, Eq. (3)] The explanation of the two scenarios is confusing. The text says 'if Di ≤ Dthreshold, the reward value will be equal to the first term,' but the expression has no first term; the term is simply 0. Rephrase to clarify that no penalty is incurred when Di ≤ Dthreshold.
[Section III-D] DDPG exploration is described as epsilon-greedy. DDPG typically uses continuous noise (e.g., Ornstein-Uhlenbeck) for exploration; epsilon-greedy over a continuous action space is not defined as written. Clarify the exploration mechanism used.
[Section IV, Table II] The learning rate is reported as a range [0.001, 0.2]. Give the specific learning rate used in the experiments, along with any scheduling or decay.
[Section III-D / IV] The What-if engine is only described as providing 'a divergent set of environmental states.' Specify how these states are generated and how they affect training or evaluation.
[Section IV] There are language errors: 'the most successive method' should be 'the most successful method' and 'continuos action set' should be 'continuous action set.' Also, the conclusion says '30% energy preservation' while Section IV says '30% reduction in energy consumption'; align the terminology.
[Section II] The statement that existing strategies 'mostly rely on discrete actions' is too broad, since [17] uses PPO, which can handle continuous action spaces. Consider softening or providing evidence.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; self-citations are background references and the reward-objective overlap is standard RL practice.

full rationale

The paper's central claims (37% latency improvement, 30% energy reduction) are empirical simulation outcomes, not derived predictions. The RL reward function directly includes battery level and a delay penalty, but this is the intended objective design: the agent is trained to minimize energy consumption and latency, and then evaluated on those same metrics. This is standard reinforcement learning practice, not a circular derivation. The consecutive-transmission penalty in Eq. (4) is identically zero because Eq. (1) enforces a'_i >= a_i + Delta_tmin, making the penalty term always zero; this is a modeling bug, not a circular step. The paper cites several prior works by the same authors (e.g., [11], [12], [14], [16], [21]) but these are background references or baselines; the proposed framework and its simulation are self-contained and do not reduce to a self-citation chain. The lack of an explicit mapping from action a_i to delay D_i and battery level B_i is a reproducibility and correctness concern, not evidence of circularity. Therefore, no step in the derivation chain reduces to its own inputs by construction, and the circularity score is low.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The main unstated inputs are the simulation's energy model, delay model, and DDPG hyperparameters. The paper contributes an architecture, but the quantitative claims depend on modeling choices that are not supplied, so a reader cannot audit the 30% and 37% numbers.

free parameters (6)

delta_t_min (minimum interval between consecutive transmissions)
Chosen by hand; not listed in Table II. It defines the action constraint in Eq. (1) and the always-zero penalty in Eq. (4), and its value changes the feasible action set.
lambda (consecutive transmission penalty factor) = 0.3
Set manually in Table II. Scales a penalty that is always zero, so its exact value has no effect on the trained rewards.
D_threshold (delay threshold in reward) = 180 ms
Set manually in Table II. Determines when the timeliness penalty activates and directly shapes the reported latency metric.
Learning rate = not reported; range [0.001, 0.2]
Table II lists only a range. The actual learning rate used in training is not given, so the DDPG result cannot be reproduced.
Discount factor = 0.8
Chosen in Table II. Affects long-term reward weighting and the final policy.
Batch size = 256
Chosen in Table II. Part of the DDPG training setup; its effect on results is unreported.

assumptions (5)

domain assumption DDPG converges to a policy that maximizes the formulated reward in this environment
The paper applies DDPG without convergence guarantees, learning-rate tuning, or replication across seeds; Section IV simply reports a single outcome.
domain assumption Battery level after an action, B_i(a_i), is computable and decreases monotonically with transmissions
Used in Eq. (2) and the energy evaluation; no battery discharge model or device energy profile is given in Section IV.
domain assumption Communication delay D_i is defined, measurable, and affected by the chosen action a_i
Used in Eq. (3) to compute R_timeliness; the paper never defines D_i, how it is simulated, or how a_i changes it.
domain assumption The Kaggle datasets [19]-[21] are representative of smart-city sensor traffic for air quality, transport, and smart farming
The simulator is said to use these datasets (Section IV), but no preprocessing or link to the simulation is described.
domain assumption The 95th-percentile latency improvement is attributable to the DT overlay and RL scheduler rather than to protocol or baseline mismatch
Baselines from [11],[13],[17],[18] target topology discovery, protocol adaptation, and offloading, not DT data-update scheduling; no ablation isolates the proposed components.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Digital Twin-Guided Energy Management over Real-Time Pub/Sub Protocol in 6G Smart Cities." pith.science (2026). https://pith.science/paper/FB5PGRAM

@misc{pith2026250818516,
  author       = {Pith},
  title        = {Pith review of: Digital Twin-Guided Energy Management over Real-Time Pub/Sub Protocol in 6G Smart Cities},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FB5PGRAM}},
  note         = {Machine review of arXiv:2508.18516}
}

read the original abstract

Although the emergence of 6G IoT networks has accelerated the deployment of enhanced smart city services, the resource limitations of IoT devices remain as a significant problem. Given this limitation, meeting the low-latency service requirement of 6G networks becomes even more challenging. However, existing 6G IoT management strategies lack real-time operation and mostly rely on discrete actions, which are insufficient to optimise energy consumption. To address these, in this study, we propose a Digital Twin (DT)-guided energy management framework to jointly handle the low latency and energy efficiency challenges in 6G IoT networks. In this framework, we provide the twin models through a distributed overlay network and handle the dynamic updates between the data layer and the upper layers of the DT over the Real-Time Publish Subscribe (RTPS) protocol. We also design a Reinforcement Learning (RL) engine with a novel formulated reward function to provide optimal data update times for each of the IoT devices. The RL engine receives a diverse set of environment states from the What-if engine and runs Deep Deterministic Policy Gradient (DDPG) to output continuous actions to the IoT devices. Based on our simulation results, we observe that the proposed framework achieves a 37% improvement in 95th percentile latency and a 30% reduction in energy consumption compared to the existing literature.

Figures

Figures reproduced from arXiv: 2508.18516 by the authors.

**Figure 2.** Management Layer, Twin Layer, and Service Layer interactions over RTPS. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** 95th percentile delay vs increasing volume of data communicated for [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Energy consumption vs number of publishers with 3-service smart [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

[11]

Ai-driven partial topology discovery algorithm for broadband networks,

K. Duran, B. Karanlik, and B. Canberk, “Ai-driven partial topology discovery algorithm for broadband networks,” in2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC) , 2021, pp. 1–6

work page 2021
[13]

Protocol-adaptive strategies for wireless mesh smart city net- works,

L. Mamatas, V . Demiroglou, S. Kalafatidis, S. Skaperas, and V . Tsaous- sidis, “Protocol-adaptive strategies for wireless mesh smart city net- works,” IEEE Network, vol. 37, no. 2, pp. 136–143, 2023

work page 2023
[17]

Aoi-aware energy control and computation offloading for industrial iot,

J. Huang, H. Gao, S. Wan, and Y . Chen, “Aoi-aware energy control and computation offloading for industrial iot,” Future Generation Computer Systems, vol. 139, pp. 29–37, 2023

work page 2023
[18]

A hybrid approach for latency and battery lifetime optimization in iot devices through offloading and cnn learning,

A. Heidari, N. J. Navimipour, M. A. J. Jamali, and S. Akbarpour, “A hybrid approach for latency and battery lifetime optimization in iot devices through offloading and cnn learning,” Sustainable Computing: Informatics and Systems , vol. 39, p. 100899, 2023

work page 2023
[1]

Energy optimization for green communication in iot using harris hawks optimization,

K. Dev, P. K. R. Maddikunta, T. R. Gadekallu, S. Bhattacharya, P. Hegde, and S. Singh, “Energy optimization for green communication in iot using harris hawks optimization,” IEEE Transactions on Green Communications and Networking , vol. 6, no. 2, pp. 685–694, 2022

work page 2022
[2]

Software defined wireless network testbed using raspberry pi of switches with routing add-on,

M. Ariman, G. Sec ¸inti, M. Erel, and B. Canberk, “Software defined wireless network testbed using raspberry pi of switches with routing add-on,” in 2015 IEEE Conference on Network Function Virtualization and Software Defined Network (NFV-SDN) , 2015, pp. 20–21

work page 2015
[3]

Fsc: Two-scale ai-driven fair sensitivity control for 802.11ax networks,

E. Ak and B. Canberk, “Fsc: Two-scale ai-driven fair sensitivity control for 802.11ax networks,” in GLOBECOM 2020 - 2020 IEEE Global Communications Conference, 2020, pp. 1–6

work page 2020
[4]

Networking architectures and protocols for smart city systems,

I. Jawhar, N. Mohamed, and J. Al-Jaroodi, “Networking architectures and protocols for smart city systems,” Journal of internet services and applications , vol. 9, no. 1, 2018. [Online]. Available: http://dx.doi.org/10.1186/s13174-018-0097-0

work page doi:10.1186/s13174-018-0097-0 2018

Show all 21 references

[5]

A decentralized framework for serverless edge computing in the internet of things,

C. Cicconetti, M. Conti, and A. Passarella, “A decentralized framework for serverless edge computing in the internet of things,” IEEE Transac- tions on Network and Service Management , vol. 18, no. 2, pp. 2166– 2180, 2021

2021
[6]

A survey of iot management protocols and frameworks,

S. Sinche, D. Raposo, N. Armando, A. Rodrigues, F. Boavida, V . Pereira, and J. S. Silva, “A survey of iot management protocols and frameworks,” IEEE Communications Surveys & Tutorials , vol. 22, no. 2, pp. 1168– 1190, 2020

2020
[7]

Predictive spectrum decision mechanisms in cognitive radio networks,

G. S. Uyanik, B. Canberk, and S. Oktug, “Predictive spectrum decision mechanisms in cognitive radio networks,” in 2012 IEEE Globecom Workshops, 2012, pp. 943–947

2012
[8]

Sdoff: A software- defined offloading controller for heterogeneous networks,

Z. Arslan, M. Erel, Y . ¨Ozcevik, and B. Canberk, “Sdoff: A software- defined offloading controller for heterogeneous networks,” in 2014 IEEE Wireless Communications and Networking Conference (WCNC) , 2014, pp. 2827–2832

2014
[9]

Stress-testing mqtt brokers: A comparative analysis of performance measurements,

B. Mishra, B. Mishra, and A. Kertesz, “Stress-testing mqtt brokers: A comparative analysis of performance measurements,” Energies, vol. 14, no. 18, 2021. [Online]. Available: https://www.mdpi.com/1996- 1073/14/18/5817

2021
[10]

A novel energy- conscious access point (eap) system with cross-layer design in wi-fi networks for reliable iot services,

S. Lee, H. Choi, T. Kim, H.-S. Park, and J. K. Choi, “A novel energy- conscious access point (eap) system with cross-layer design in wi-fi networks for reliable iot services,” IEEE Access , vol. 10, pp. 61 228– 61 248, 2022

2022
[12]

Digital twin enriched green topology discovery for next generation core networks,

K. Duran and B. Canberk, “Digital twin enriched green topology discovery for next generation core networks,” IEEE Transactions on Green Communications and Networking , vol. 7, no. 4, pp. 1946–1956, 2023

1946
[14]

Gentwin: Generative ai-powered digital twinning for adaptive management in iot networks,

K. Duran, H. Shin, T. Q. Duong, and B. Canberk, “Gentwin: Generative ai-powered digital twinning for adaptive management in iot networks,” IEEE Transactions on Cognitive Communications and Networking , pp. 1–1, 2025

2025
[15]

Intelligent edge computing for iot-based energy management in smart cities,

Y . Liu, C. Yang, L. Jiang, S. Xie, and Y . Zhang, “Intelligent edge computing for iot-based energy management in smart cities,” IEEE Network, vol. 33, no. 2, pp. 111–117, 2019

2019
[16]

Intelligent digital twin communication framework for addressing accuracy and timeliness tradeoff in resource-constrained networks,

L. V . Cakir, C. J. Thomson, M. ¨Ozdem, B. Canberk, V .-L. Nguyen, and T. Q. Duong, “Intelligent digital twin communication framework for addressing accuracy and timeliness tradeoff in resource-constrained networks,” IEEE Transactions on Cognitive Communications and Net- worki...

2024
[19]

Smoke detection dataset,

S. Blattmann, “Smoke detection dataset,” 2022. [Online]. Avail- able: https://www.kaggle.com/datasets/deepcontractor/smoke-detection- dataset/data

2022
[20]

A simple dataset of aquaponic fish pond iot,

B. Siswanto and B. Mardiyana, “A simple dataset of aquaponic fish pond iot,” 2023. [Online]. Available: https://www.kaggle.com/dsv/5225463

2023
[21]

Digital twin-empowered green mobility management in next-gen trans- portation networks,

K. Duran, L. V . Cakir, A. Fonzone, T. Q. Duong, and B. Canberk, “Digital twin-empowered green mobility management in next-gen trans- portation networks,” IEEE Open Journal of Vehicular Technology, vol. 5, pp. 1650–1662, 2024

2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [11]

Ai-driven partial topology discovery algorithm for broadband networks,

K. Duran, B. Karanlik, and B. Canberk, “Ai-driven partial topology discovery algorithm for broadband networks,” in2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC) , 2021, pp. 1–6

work page 2021

[2] [13]

Protocol-adaptive strategies for wireless mesh smart city net- works,

L. Mamatas, V . Demiroglou, S. Kalafatidis, S. Skaperas, and V . Tsaous- sidis, “Protocol-adaptive strategies for wireless mesh smart city net- works,” IEEE Network, vol. 37, no. 2, pp. 136–143, 2023

work page 2023

[3] [17]

Aoi-aware energy control and computation offloading for industrial iot,

J. Huang, H. Gao, S. Wan, and Y . Chen, “Aoi-aware energy control and computation offloading for industrial iot,” Future Generation Computer Systems, vol. 139, pp. 29–37, 2023

work page 2023

[4] [18]

A hybrid approach for latency and battery lifetime optimization in iot devices through offloading and cnn learning,

A. Heidari, N. J. Navimipour, M. A. J. Jamali, and S. Akbarpour, “A hybrid approach for latency and battery lifetime optimization in iot devices through offloading and cnn learning,” Sustainable Computing: Informatics and Systems , vol. 39, p. 100899, 2023

work page 2023

[5] [1]

Energy optimization for green communication in iot using harris hawks optimization,

K. Dev, P. K. R. Maddikunta, T. R. Gadekallu, S. Bhattacharya, P. Hegde, and S. Singh, “Energy optimization for green communication in iot using harris hawks optimization,” IEEE Transactions on Green Communications and Networking , vol. 6, no. 2, pp. 685–694, 2022

work page 2022

[6] [2]

Software defined wireless network testbed using raspberry pi of switches with routing add-on,

M. Ariman, G. Sec ¸inti, M. Erel, and B. Canberk, “Software defined wireless network testbed using raspberry pi of switches with routing add-on,” in 2015 IEEE Conference on Network Function Virtualization and Software Defined Network (NFV-SDN) , 2015, pp. 20–21

work page 2015

[7] [3]

Fsc: Two-scale ai-driven fair sensitivity control for 802.11ax networks,

E. Ak and B. Canberk, “Fsc: Two-scale ai-driven fair sensitivity control for 802.11ax networks,” in GLOBECOM 2020 - 2020 IEEE Global Communications Conference, 2020, pp. 1–6

work page 2020

[8] [4]

Networking architectures and protocols for smart city systems,

I. Jawhar, N. Mohamed, and J. Al-Jaroodi, “Networking architectures and protocols for smart city systems,” Journal of internet services and applications , vol. 9, no. 1, 2018. [Online]. Available: http://dx.doi.org/10.1186/s13174-018-0097-0

work page doi:10.1186/s13174-018-0097-0 2018