Pith. sign in

REVIEW 4 major objections 5 minor 40 references

Prediction and Communication Co-design for Ultra-Reliable and Low-Latency Communications

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper argues that sending predicted future states early, with an optimized prediction horizon, lets ultra-reliable low-latency links meet delay and reliability targets that no-prediction transmission cannot satisfy.

desk verdict The co-design idea is genuinely new and plausible, but the reported zero-latency result is infeasible under the paper's own delay definitions and the prediction-tail formula misses a factor of two. read the letter →

arxiv 1909.05787 v1 pith:RVJUE3EU submitted 2019-09-06 eess.SP cs.ITcs.NIcs.SYeess.SYmath.IT

Zhanwei Hou , Changyang She , Yonghui Li , Zhuo Li , Branka Vucetic This is my paper · ORCID
classification eess.SPcs.ITcs.NIcs.SYeess.SYmath.IT
keywords ultra-reliableandlow-latencycommunicationspredictioncommunicationco-designdelay-reliabilitytradeoffhorizonshortblocklengtheffectivebandwidthK-repetitionTactileInternet
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 tries to establish that ultra-reliable and low-latency communications can be achieved without sacrificing user experience by having each mobile device predict its future states and transmit those predictions ahead of time, then jointly optimizing the prediction horizon with queueing delay, transmission delay, and bandwidth. The load-bearing relation is that user experienced delay equals communication delay minus prediction horizon, so sending future states early converts network latency into a prediction-reliability problem. The authors derive closed forms for prediction error probability under a linear Gaussian state model, queueing delay violation probability via effective bandwidth, and decoding error probability over short-blocklength fading channels, then minimize the bandwidth per device subject to a total error budget. In a single-user simulation, the co-design reaches an overall error probability of $6.52\times10^{-6}$ at zero experienced delay with a prediction horizon of 26.8 ms, while the no-prediction baseline cannot satisfy even a 10 ms delay bound. A tactile-Internet remote-control experiment confirms the predicted monotonic growth of prediction error with horizon, although measured error probabilities are higher than ideal simulation because velocity and acceleration must be estimated from position data.

What carries the argument

The central object is the experienced-delay identity $D_e^n=D_c^n-T_p^n$, with $D_c^n=D_q^n+D_t^n+D_d^n+D_r^n$. It couples the two sides: raising the prediction horizon $T_p$ shrinks user experienced delay but increases the prediction error variance $\rho_{n,j}^2(T_p)=\sigma_j^2+\sum_{i=1}^{T_p-1}\sum_{m=1}^{F}(\phi_{n,j,m,T_p-i})^2\sigma_m^2$, while the extra communication delay lowers $\varepsilon_q$ and $\varepsilon_t$. The optimization balances the three error terms by enforcing $\varepsilon_q+\varepsilon_t+\varepsilon_p\le\varepsilon_{\max}$, with $\varepsilon_q$ from effective bandwidth, $\varepsilon_t=(\bar\varepsilon_\tau)^{K_n}$ from K-repetition of a short-blocklength code, and the near-optimal simplification $\varepsilon_q=\varepsilon_t$ plus a bisection search on $T_p$.

What would settle it

On a real device obeying the assumed kinematics, record $T_s$-spaced positions for a long session, form the open-loop prediction $\hat X_n(k+T_p)=\Phi^{T_p}X_n(k)$ at the optimized horizon (e.g., $T_p=26.8$ ms), and count how often $|e_j^n(k)|>\delta_j$ with $\delta_j=0.1$ m. If the measured prediction error probability is materially above the Gaussian closed-form value, the optimized allocation no longer guarantees the claimed reliability in practice.

Watch

Extended reading notes

Core claim

The paper's central claim is that prediction and communication should be designed as one resource-allocation problem, not separately. Given a state transition model $X_n(k+1)=\Phi_n X_n(k)+W_n(k)$, the open-loop predictor $\hat X_n(k+T_p)=(\Phi_n)^{T_p}X_n(k)$ produces a Gaussian prediction error whose variance grows with horizon, so the prediction error probability $\varepsilon_p$ is a closed-form increasing function of $T_p$. The other two error components—the queueing delay violation probability $\varepsilon_q$ and the packet loss probability $\varepsilon_t$—decrease when the communication system is allowed more delay, which is exactly what a longer prediction horizon buys. The paper therefore maximizes the number of supported devices by minimizing per-user bandwidth over $T_p, D_q, D_t$ under $D_q+D_t+D_d+D_r-T_p\le D_{\max}$ and $\varepsilon_q+\varepsilon_t+\varepsilon_p\le\varepsilon_{\max}$, using a near-optimal binary-search algorithm whose gap to the optimum is bounded by the optimal error probability itself. The numerical headline is that zero experienced delay becomes feasible at error probability $6.52\times10^{-6}$, a point on the delay–reliability curve that no-prediction operation cannot reach.

Load-bearing premise

The whole prediction-error calculation rests on a linear, time-invariant Gaussian state model with a known constant transition matrix; if real device dynamics are nonlinear or non-Gaussian, the closed-form error probability no longer follows and the optimized allocation may miss its reliability target.

Editorial extensions

If this is right

  • With prediction enabled, the delay budget can be larger than the end-to-end bound, so backhaul and core network delays stop being a hard barrier to ultra-low user experienced delay.
  • At the simulated single-user settings, zero experienced delay is reported with overall error $6.52\times10^{-6}$, whereas no-prediction cannot satisfy a 10 ms delay bound.
  • For a fixed reliability of $10^{-5}$, prediction co-design cuts user experienced delay by roughly 23 ms compared with no prediction at the same bandwidth and antennas.
  • In multi-user deployment, increasing AP antennas is an efficient substitute for bandwidth: with 1 MHz and known fading distribution, going from 32 to 64 antennas raises supported devices from about 2 to about 10; in worst-case design it saves roughly 75% of bandwidth.
  • The per-user bandwidth search adds only a linear factor in the number of users, so the design stays feasible for larger URLLC deployments.

Reading between the lines

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

  • Because the optimization only needs a prediction-error distribution, the same co-design should work with data-driven predictors that output an estimated error variance, as long as that estimate reflects the true tail probability; the Gaussian assumption is a convenience, not the core mechanism.
  • The experienced-delay identity suggests a natural per-device adaptation: devices with smooth, predictable motion get longer horizons and lower bandwidth, while erratic devices are allocated more resources, making predictability a schedulable resource.
  • A direct experiment on human teleoperation traces could test the boundary of the claim: the horizon $T_p$ that meets the reliability target in ideal Gaussian simulation should be re-tuned to the empirical error distribution before deployment.
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.

Referee Report

4 major / 5 minor

Summary. This paper proposes a prediction and communication co-design framework for URLLC, in which each mobile device predicts its future state and transmits the prediction ahead of time, so that the user experienced delay can be reduced at the cost of prediction errors. The authors derive closed-form expressions for the prediction error probability, queueing delay violation probability, and short-packet decoding error probability, formulate an optimization problem to maximize the number of supported devices under delay and reliability constraints, and propose a per-user bandwidth minimization algorithm with a near-optimal prediction horizon. Numerical results in Section VI-A claim that the proposed method achieves an overall error probability of 6.52e-6 at zero experienced delay in a single-user scenario, and an experiment with a real tactile device is used to validate the mobility prediction concept.

Significance. If the internal consistency issues are resolved, the co-design idea is relevant and timely: it couples the prediction horizon with communication resource allocation, which is a useful angle for URLLC system design. The paper's strengths include closed-form characterizations of the three error components, the decomposition of the network optimization into per-user problems with linear search complexity, and a real-device experiment that goes beyond pure simulation. However, the numerical verification in Section VI evaluates the paper's own analytical expressions rather than an independent Monte Carlo simulation, and several load-bearing equations are internally inconsistent. The headline zero-latency result and the reported reliability value therefore need to be corrected and re-established before the paper can be accepted.

major comments (4)
  1. [Sec. IV-B, Eq. (13)] Equation (13) evaluates Pr{|e_j^n(k+T_p^n)|>delta_j} as Psi(-delta_j/rho_{n,j}(T_p^n)), where Psi is the standard Gaussian CDF. For a zero-mean Gaussian random variable the two-sided tail is 2 Psi(-delta/rho), so the expression misses the upper tail. In the ultra-reliable regime Psi(-delta/rho) is very small and the underestimation approaches a factor of two. This error propagates into Eqs. (14) and (23f) and into the reported overall error probability 6.52e-6 in Section VI-A; correcting just the factor of two would put that operating point above the 10^-5 target. The authors should either correct Eq. (13) and recalculate all numerical results, or justify a different definition of Psi in the text.
  2. [Sec. III-B, Sec. V-A, Sec. VI-A, Eqs. (2), (23b), (27)] Constraint (23b) includes the decoding delay D_d^n=kappa D_t^n with kappa>0, but the optimality condition in Eq. (27) is written as D_q^n+D_t^n+D_r^n-T_p^n=D_max, and the text around Fig. 3 equates D_q+D_t=T_p with zero experienced delay. Both statements drop D_d and D_r. Using the paper's own reported numbers D_q=14.3 ms, D_t=12.5 ms, T_p=26.8 ms, and D_r=10 ms, Eq. (2) gives D_e=14.3+12.5+kappa*12.5+10-26.8=10+kappa*12.5 ms>0. Even if kappa were zero, D_e=10 ms, not 0. The headline 'zero-latency' single-user result therefore does not satisfy the delay constraint (23b) with D_max=0, and the discussion of Fig. 3 must be revised.
  3. [Sec. V-B and Appendix D] Proposition 1 and the binary-search solution use Eq. (27) with the D_d term omitted. Under the true delay constraint D_q+(1+kappa)D_t+D_r-T_p=D_max, the relation D_q+D_t=D_max+T_p-D_r used in Appendix D is invalid, and the monotonicity argument for epsilon_q and epsilon_t with respect to T_p needs to be re-derived. Consequently the proposed algorithm solves a relaxed problem that overstates the feasible region, so the minimal bandwidth and supported-device counts in Section VI are not the solution to the stated optimization problem.
  4. [Sec. VI-A and Table III] The reported single-user operating point is internally inconsistent with the K-repetition model. Table III sets D_tau=0.5 ms, while the text reports K=5 and D_t=12.5 ms. Since the paper defines D_t=K D_tau before Eq. (19), these values give D_t=2.5 ms, not 12.5 ms. Either the parameter D_tau or the reported D_t must be corrected; as written, the numerical results cannot be reproduced.
minor comments (5)
  1. [Table II] The pseudocode uses 'while |...| < epsilon_max do' for the binary search; a binary search should continue while the difference is above the tolerance. As written, the loop can exit immediately when the initial estimate is not already within the tolerance.
  2. [Fig. 3 and Sec. VI-A] Fig. 3 labels the communication error probability as epsilon_c, while the text and equations use epsilon_t; please unify the notation.
  3. [Sec. VI-B] The experiment estimates velocity and acceleration by finite differences of the location sequence, which is a different state model from the constant-Phi model in Eqs. (6)-(9). The paper should state explicitly that the experiment is a qualitative validation of the prediction concept, not a direct validation of the closed-form probability expressions.
  4. [Tables IV and V] The layout of Table V is confusing because the two nTs columns correspond to different values of delta; please separate the two conditions more clearly or use a sub-table format.
  5. [Eq. (13)] The notation Psi should be defined unambiguously as the standard Gaussian CDF, and the resulting two-sided probability should be written explicitly so that readers do not mistake Psi(-delta/rho) for the tail probability.

Circularity Check

1 steps flagged · score 4.0 of 10

The optimization is not globally circular, but the proof of Proposition 1 is circular because it assumes Eq. (27) before that equality is established; self-citations support standard components and are not counted as circular.

  1. other [Appendix D, proof of Proposition 1 (Section V-B); Eq. (27) in Section V-B]
    "Proof. According to (27), we have Dq n +Dt n =Dmax +T p n − Dr n. To prove this proposition, we need to prove that εq n or εt n decreases with Dq n +Dt n."

    Proposition 1 is the load-bearing monotonicity claim used to reduce the optimization to a one-variable search over the prediction horizon Tp. Its proof begins by invoking Eq. (27), the equality Dq+Dt+Dr−Tp=Dmax. However, Eq. (27) is introduced in the main text immediately after Proposition 1 as a consequence of Proposition 1 plus Lemma 1: 'together with Proposition 1, the optimal solution is obtained when the equality in (27) holds.' Thus the proof of the monotonicity result assumes the very optimality equality whose derivation depends on that monotonicity. The subsequent variable-reduction step, the 'near optimal' claim, and Proposition 2 therefore rest on a circular proof as written.

full rationale

The central co-design problem is not a fit renamed as a prediction: the closed-form expressions for the prediction error probability, queueing-delay violation probability, and decoding error probability are derived from explicit models (linear Gaussian state evolution, effective bandwidth, and finite-blocklength normal approximation) rather than from the target result. The self-citations to [17], [21], and [39] support standard or previously published approximations and are peer-reviewed, so under the stated rules they are real evidence rather than circularity. The main circular step is in the proof of Proposition 1, which uses Eq. (27) before Eq. (27) is justified, leaving the optimality argument for the prediction-horizon search without an independent proof. Separately, the numerical verification is partly self-referential because the simulation curves are generated from the same derived expressions being optimized, so the reported 6.52e-6 value is not an independent Monte Carlo confirmation. The real-device experiment does provide an external check of the prediction-error trend, but it is not used to validate the full optimized delay allocation. I did not count the internal-consistency problems (Dd dropped in Eq. (27), Dr omitted in the Fig. 3 zero-delay discussion, and the one-sided tail in Eq. (13)) as circularity because they are correctness concerns rather than reductions of the output to the input; they should be weighed separately in a correctness review.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central analysis assumes a known linear-Gaussian state model, a specific open-loop predictor, standard effective bandwidth for Poisson traffic, and short-blocklength approximations from prior work. The only genuinely ad hoc modeling choice introduced by the paper is the equality eps_q = eps_t used for tractability. No new physical entities are postulated.

free parameters (3)
  • Acceleration transition noise sigma_w = 0.01 m/s^2
    Set by hand for the simulation in Section VI-A. The prediction error variance in Eq. (12) and hence eps_p in Eq. (14) scale with this value; no measurement or fitting procedure is given.
  • Prediction threshold delta_l = 0.1 m
    Chosen as the required accuracy for the one-dimensional movement example. The prediction error probability in Eq. (14) depends directly on this threshold.
  • Worst-case shadowing margin S_w = -34.1 dB
    Chosen so that Pr{S_n <= S_w} = 10^-5 in Section VI-A. This margin determines the SNR used in the reliability calculations.
assumptions (6)
  • domain assumption The device state evolves as Xn(k+1)=Phi Xn(k)+Wn(k) with known constant Phi and Gaussian transition noise.
    Used to derive the prediction error variance in Eq. (12) and the prediction error probability in Eq. (14). Remark 2 acknowledges this does not cover nonlinear systems.
  • domain assumption The predictor is open-loop: Xhat(k+Tp)=Phi^{Tp} Xn(k).
    This specific predictor choice produces the Gaussian prediction error. Other predictors, for example with measurement updates, would give different reliability values.
  • domain assumption The short-blocklength normal approximation [15] and the closed-form approximation [39] accurately describe decoding error over Rayleigh fading.
    Used in Eqs. (19)-(21) to obtain the packet loss probability. The paper relies on the cited accuracy of these approximations rather than verifying them in this setting.
  • domain assumption Packet arrivals follow a Poisson or switched Poisson process with known average rate lambda_n, so effective bandwidth applies.
    Used to derive the queueing delay violation probability in Eq. (17). The paper notes real Tactile Internet traffic is bursty and uses a two-state SPP model.
  • ad hoc to paper Setting eps_q = eps_t is a tractability simplification with negligible performance loss.
    Imposed in Section V-B before Proposition 1. The paper cites [17] rather than proving that this equality is near-optimal in the present prediction and communication setting.
  • standard math For small error probabilities, the overall reliability is approximated by eps_q + eps_t + eps_p.
    Eq. (5) replaces the product form with a sum. This is valid only when all three terms are far below 1, which is consistent with the URLLC targets considered.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Prediction and Communication Co-design for Ultra-Reliable and Low-Latency Communications." pith.science (2026). https://pith.science/paper/RVJUE3EU

@misc{pith2026190905787,
  author       = {Pith},
  title        = {Pith review of: Prediction and Communication Co-design for Ultra-Reliable and Low-Latency Communications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RVJUE3EU}},
  note         = {Machine review of arXiv:1909.05787}
}
read the original abstract

Ultra-reliable and low-latency communications (URLLC) are considered as one of three new application scenarios in the fifth generation cellular networks. In this work, we aim to reduce the user experienced delay through prediction and communication co-design, where each mobile device predicts its future states and sends them to a data center in advance. Since predictions are not error-free, we consider prediction errors and packet losses in communications when evaluating the reliability of the system. Then, we formulate an optimization problem that maximizes the number of URLLC services supported by the system by optimizing time and frequency resources and the prediction horizon. Simulation results verify the effectiveness of the proposed method, and show that the tradeoff between user experienced delay and reliability can be improved significantly via prediction and communication co-design. Furthermore, we carried out an experiment on the remote control in a virtual factory, and validated our concept on prediction and communication co-design with the practical mobility data generated by a real tactile device.

Figures

Figures reproduced from arXiv: 1909.05787 by the authors.

Figure 1
Figure 1. Illustration of network structure. and thus can enable haptic communications. The packets generated by each device may include different features, such as the location, velocity and acceleration of a device in remote driving or industrial automation, or the force and torques in Tactile Internet. The receiver can be deployed at a mobile edge computing (MEC) server or a cloud center. In our framework, we consider a ge… view at source ↗
Figure 2
Figure 2. Illustration of prediction and communication co-de [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Joint optimization of predictions and communicatio [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Comparison of reliability-delay tradeoff curves be [PITH_FULL_IMAGE:figures/full_fig_p022_4.png]
Figure 5
Figure 5. Figure 5: Pr{ PN n=1 Bn > Bmax} v.s. the number of devices when the distribution of large-scale fading of devices is known. Since the large-scale fading of devices are random variables in the first scenario, the sum of the required bandwidth is also a random variable. In [PITH_…
Figure 6
Figure 6. Figure 6: Required total bandwidth v.s. number of devices when [PITH_FULL_IMAGE:figures/full_fig_p024_6.png]
Figure 7
Figure 7. Figure 7: Experiment to obtain real movement data in Tactile In [PITH_FULL_IMAGE:figures/full_fig_p025_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 39 canonical work pages

  1. [19]

    Minimizing wireless resource consumption for packetize d predictive control in real-time cyber physical systems,

    X. Tong, G. Zhao, M. A. Imran et al. , “Minimizing wireless resource consumption for packetize d predictive control in real-time cyber physical systems,” in Proc. IEEE ICC W orkshops, May 2018

    work page 2018

  2. [20]

    5G-enabled tactile internet,

    M. Simsek, A. Aijaz, M. Dohler, J. Sachs, and G. Fettweis , “5G-enabled tactile internet,” IEEE J. Sel. Areas Commun. , vol. 34, no. 3, pp. 460–473, Mar. 2016

    work page 2016

  3. [17]

    Cross-layer optimiza tion for ultra-reliable and low-latency radio access netwo rks,

    C. She, C. Yang, and T. Q. S. Quek, “Cross-layer optimiza tion for ultra-reliable and low-latency radio access netwo rks,” IEEE Wireless Commun. , vol. 17, no. 1, pp. 127–141, 2018

    work page 2018

  4. [18]

    Reliability ana lysis of uplink grant-free transmission over shared resources,

    G. Berardinelli, N. H. Mahmood, R. Abreu, T. Jacobsen, K . Pedersen, I. Z. Kov´ acs, and P . Mogensen, “Reliability ana lysis of uplink grant-free transmission over shared resources,” IEEE Access , vol. 6, pp. 23 602–23 611, 2018

    work page 2018

  5. [1]

    Ultra-reliable and low-latency communications: prediction and communicatio n co-design,

    Z. Hou, C. She, Y . Li, and B. Vucetic, “Ultra-reliable and low-latency communications: prediction and communicatio n co-design,” in Proc. IEEE ICC , 2019

    work page 2019

  6. [2]

    Study on scenarios and requir ements for next generation access technologies,

    3GPP TSG RAN TR38.913 R14, “Study on scenarios and requir ements for next generation access technologies,” Jun. 2017

    2017

  7. [3]

    Augmenting self -driving with remote control: Challenges and directions,

    L. Kang, W. Zhao, B. Qi, and S. Banerjee, “Augmenting self -driving with remote control: Challenges and directions,” in Proc. ACM Mobile Computing Systems & Applications , 2018, pp. 19–24

    work page 2018

  8. [4]

    Toward re al-time control in future wireless networks: communicatio n- control co-design,

    G. Zhao, M. A. Imran, Z. Pang, Z. Chen, and L. Li, “Toward re al-time control in future wireless networks: communicatio n- control co-design,” IEEE Commun. Mag. , vol. 57, no. 2, pp. 138–144, 2019

    work page 2019

Show all 40 references
  1. [5]

    The Tactile Internet: applications & ch allenges,

    G. P . Fettweis, “The Tactile Internet: applications & ch allenges,” IEEE V eh. Technol. Mag., vol. 9, no. 1, pp. 64–70, Mar. 2014

    2014

  2. [6]

    Study on new radio (NR) access tec hnology; physical layer aspects (release 14),

    3GPP TR 38.802 V2.0.0, “Study on new radio (NR) access tec hnology; physical layer aspects (release 14),” 2017

    2017

  3. [7]

    Radio resource managem ent for ultra-reliable and low-latency communications,

    C. She, C. Yang, and T. Q. S. Quek, “Radio resource managem ent for ultra-reliable and low-latency communications,” IEEE Commun. Mag. , vol. 55, no. 6, pp. 72–78, 2017

    2017

  4. [8]

    Ultra-reliable and low latency communication in mmwave- enabled massive MIMO networks,

    T. K. Vu, C.-F. Liu, M. Bennis et al. , “Ultra-reliable and low latency communication in mmwave- enabled massive MIMO networks,” IEEE Commun. Letters , vol. 21, no. 9, pp. 2041–2044, Sep. 2017

    2017

  5. [9]

    Ultra-reliable l ow latency communication using interface diversity,

    J. J. Nielsen, R. Liu, and P . Popovski, “Ultra-reliable l ow latency communication using interface diversity,” IEEE Trans. on Commun. , vol. 66, no. 3, pp. 1322–1334, Mar. 2018

    2018

  6. [10]

    5G New Radio: unveiling the essentials of the next generat ion wireless access technology,

    X. Lin, J. Li, R. Baldemair et al. , “5G New Radio: unveiling the essentials of the next generat ion wireless access technology,” 2018. [Online]. Available: https://arxiv.o rg/abs/1806.06898

    2018 arXiv

  7. [11]

    Latency critical IoT applications in 5G: Perspective on t he design of radio interface and network architecture,

    P . Schulz, M. Matthe, H. Klessig et al. , “Latency critical IoT applications in 5G: Perspective on t he design of radio interface and network architecture,” IEEE Commun. Mag. , vol. 55, no. 2, pp. 70–78, 2017

    2017

  8. [12]

    System level analysis of uplink grant-free transmission for URLLC,

    T. Jacobsen, R. Abreu, G. Berardinelli et al. , “System level analysis of uplink grant-free transmission for URLLC,” in Proc. IEEE Globecom W orkshops, 2017, pp. 1–6

    2017

  9. [13]

    Fundamental tradeoffs among reliability, latency and throughput in cellular networks,

    B. Soret, P . Mogensen, K. I. Pedersen, and M. C. Aguayo-T orres, “Fundamental tradeoffs among reliability, latency and throughput in cellular networks,” in Proc. IEEE Globecom W orkshops, 2014, pp. 1391–1396

    2014

  10. [14]

    5G radio network design for ultra-reliable low-latency c ommunication,

    J. Sachs, G. Wikstrom, T. Dudda et al. , “5G radio network design for ultra-reliable low-latency c ommunication,” IEEE Network, vol. 32, no. 2, pp. 24–31, 2018. September 13, 2019 DRAFT 31

    2018

  11. [15]

    Quasi-s tatic multiple-antenna fading channels at finite blockleng th,

    W. Yang, G. Durisi, T. Koch, and Y . Polyanskiy, “Quasi-s tatic multiple-antenna fading channels at finite blockleng th,” IEEE Trans. Inf. Theory , vol. 60, no. 7, pp. 4232–4265, Jul. 2014

    2014

  12. [16]

    Ultra-reliable an d low-latency wireless communication: Tail, risk and scale ,

    M. Bennis, M. Debbah, and H. V . Poor, “Ultra-reliable an d low-latency wireless communication: Tail, risk and scale ,” arXiv preprint arXiv:1801.01270 , 2018

    2018 arXiv

  13. [21]

    Burstiness aware bandwidth reservation for ultra-relia ble and low-latency communications (URLLC) in Tactile Internet,

    Z. Hou, C. She, Y . Li et al. , “Burstiness aware bandwidth reservation for ultra-relia ble and low-latency communications (URLLC) in Tactile Internet,” IEEE J. Sel. Areas Commun. , vol. 36, no. 11, pp. 2401–2410, 2018

    2018

  14. [22]

    Predictive pr e-allocation for low-latency uplink access in industrial w ireless networks,

    M. Li, X. Guan, C. Hua, C. Chen, and L. Lyu, “Predictive pr e-allocation for low-latency uplink access in industrial w ireless networks,” in Proc. IEEE INFOCOM , 2018, pp. 306–314

    2018

  15. [23]

    Fast HARQ over finite blocklength codes: A technique for low-latency reliable communication,

    B. Makki, T. Svensson, G. Caire, and M. Zorzi, “Fast HARQ over finite blocklength codes: A technique for low-latency reliable communication,” IEEE Wireless Commun. , vol. 18, no. 1, pp. 194–209, 2018

    2018

  16. [24]

    Enhanced machine learning techniques for early HARQ feed back prediction in 5G,

    N. Strodthoff, B. G¨ oktepe, T. Schierl et al. , “Enhanced machine learning techniques for early HARQ feed back prediction in 5G,” arXiv preprint arXiv:1807.10495 , 2018

    2018 arXiv

  17. [25]

    Cross-layer design for mission-critical IoT in mobile edge computing systems,

    C. She, Y . Duan, G. Zhao, T. Q. S. Quek, Y . Li, and B. Vuceti c, “Cross-layer design for mission-critical IoT in mobile edge computing systems,” IEEE Internet of Things J., early access , 2019

    2019

  18. [26]

    Human force discrimination during active arm motion for f orce feedback design,

    S. Feyzabadi, S. Straube, M. Folgheraiter et al. , “Human force discrimination during active arm motion for f orce feedback design,” IEEE Trans. Haptics , vol. 6, no. 3, pp. 309–319, 2013

    2013

  19. [27]

    S. M. Kay, Fundamentals of statistical signal processing, volume I: e stimation theory . Prentice Hall, 1993

    1993

  20. [28]

    Tracking-error model-based predictive control for mobile robots in real time,

    G. Klanˇ car and I. ˇSkrjanc, “Tracking-error model-based predictive control for mobile robots in real time,” Robotics and autonomous systems , vol. 55, no. 6, pp. 460–469, 2007

    2007

  21. [29]

    Effective bandwidth in hi gh-speed digital networks,

    C.-S. Chang and J. A. Thomas, “Effective bandwidth in hi gh-speed digital networks,” IEEE J. Sel. Areas Commun. , vol. 13, no. 6, pp. 1091–1100, 1995

    1995

  22. [30]

    Network -layer performance analysis of multihop fading channels,

    H. Al-Zubaidy, J. Liebeherr, and A. Burchard, “Network -layer performance analysis of multihop fading channels,” IEEE/ACM Trans. Netw. , vol. 24, no. 1, pp. 204–217, Feb. 2016

    2016

  23. [31]

    Soft resource reservation for low-delayed teleoperation over mobile networks,

    M. Condoluci, T. Mahmoodi, E. Steinbach, and M. Dohler, “Soft resource reservation for low-delayed teleoperation over mobile networks,” IEEE Access , vol. 5, pp. 10 445–10 455, May 2017

    2017

  24. [32]

    Delay performance of wireless communications with imperfe ct CSI and finite-length coding,

    S. Schiessl, H. Al-Zubaidy, M. Skoglund, and J. Gross, “ Delay performance of wireless communications with imperfe ct CSI and finite-length coding,” IEEE Trans. Commun. , vol. 66, no. 12, pp. 6527–6541, 2018

    2018

  25. [33]

    S hort packets over block-memoryless fading channels: Pilot-assisted or noncoherent transmission?

    J. ¨Ostman, G. Durisi, E. G. Str¨ om, M. C. Cos ¸kun, and G. Liva, “S hort packets over block-memoryless fading channels: Pilot-assisted or noncoherent transmission?” IEEE Trans. Commun. , vol. 67, no. 2, pp. 1521–1536, 2018

    2018

  26. [34]

    Optimizing pilot overhead for ultra-reliable short-packet transmission,

    M. Mousaei and B. Smida, “Optimizing pilot overhead for ultra-reliable short-packet transmission,” in Proc. IEEE ICC , 2017, pp. 1–5

    2017

  27. [35]

    How much training is need ed in multiple-antenna wireless links?

    B. Hassibi and B. M. Hochwald, “How much training is need ed in multiple-antenna wireless links?” IEEE Trans. Inf. Theory, vol. 49, no. 4, pp. 951–963, 2003. September 13, 2019 DRAFT 32

    2003

  28. [36]

    Saddlepoint approximations for Rayleigh block-fading c hannels,

    A. Lancho, J. Ostman, G. Durisi et al. , “Saddlepoint approximations for Rayleigh block-fading c hannels,” 2019. [Online]. Available: https://arxiv.org/abs/1904.10442

    2019 arXiv

  29. [37]

    Short block-length codes for ultra-reliable low latency communications,

    M. Shirvanimoghaddam, M. S. Mohammadi, R. Abbas et al. , “Short block-length codes for ultra-reliable low latency communications,” IEEE Commun. Mag. , vol. 57, no. 2, pp. 130–137, 2018

    2018

  30. [38]

    Delay analys is for wireless fading channels with finite blocklength chan nel coding,

    S. Schiessl, J. Gross, and H. Al-Zubaidy, “Delay analys is for wireless fading channels with finite blocklength chan nel coding,” in Proc. ACM MSWiM , 2015, pp. 13–22

    2015

  31. [39]

    Improving network availability of ultra-reliable and lo w-latency communications with multi-connectivity,

    C. She, Z. Chen, C. Yang et al. , “Improving network availability of ultra-reliable and lo w-latency communications with multi-connectivity,” IEEE Trans. Commun. , vol. 66, no. 11, pp. 5482–5496, Nov. 2018

    2018

  32. [40]

    Mathews and K

    J. Mathews and K. Fink, Numerical methods using MATLAB . Pearson Prentice Hall, NJ, 2004. September 13, 2019 DRAFT

    2004

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.