Power Efficient Discontinuous Reception in THz and mmWave Wireless Systems

Christopher Slezak; Sourjya Dutta; Sundar Aditya; Sundeep Rangan; Syed Hashim Ali Shah

REVIEW 4 major objections 4 minor 15 references

Power Efficient Discontinuous Reception in THz and mmWave Wireless Systems

T0 review · 4 major / 4 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 In dense mmWave and THz networks, a phone can monitor just four cells and still have a usable link 99% of the time.

desk verdict The K=4 reliability result is plausible, but the 'most power saving' half of the headline is not reproducible because TSS,0 is never specified. read the letter →

arxiv 1908.08966 v1 pith:Z7KGPDKI submitted 2019-08-23 eess.SP

Syed Hashim Ali Shah , Sundar Aditya , Sourjya Dutta , Christopher Slezak , Sundeep Rangan This is my paper · ORCID

classification eess.SP

keywords DiscontinuousreceptionDRXmillimeterwaveterahertzblockagemacro-diversitypowerconsumptionlinkreliability

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

At millimeter-wave and terahertz frequencies, a phone's receiver front end can draw close to a watt, so discontinuous reception (DRX), where the receiver is switched off between data bursts, is essential for battery life. The paper argues that any single directional link is easily blocked, so the phone must monitor several cells, and the central question is how many. It proposes a simple greedy listening-set selection and, using standard double knife-edge blockage trajectories at 28 and 140 GHz, finds that tracking four cells keeps the probability of having at least one usable link above 99% while leaving more than 85% of the DRX cycle for sleeping. If this holds on real hardware and traffic patterns, connected-mode DRX can give THz handhelds both reliability and power savings despite faster, deeper blockages.

What carries the argument

The central object is the listening set $A_n$, a subset of $K$ cells whose signals the UE measures during each DRX monitoring instance, with $T_{\text{SS}} = K T_{\text{SS,0}}$ as the awake time and $T_{\text{SS,per}} = 20$ ms as the SSB period. The cost model bounds the awake fraction as $\beta_{\text{awake}} \geq (1 - P_B) K T_{\text{SS,0}}/T_{\text{SS,per}} + P_B$, where $P_B$ is the probability that all monitored links fall below the SNR threshold $\gamma_{\min}$; this formula captures the trade-off between listening to more cells and spending more time awake. The algorithm carries the argument by keeping the best measured cell as the serving cell and triggering a full beam sweep only when every link in the listening set is blocked, which is what makes a small $K$ both reliable and power-efficient.

What would settle it

Re-run the simulation with a physically measured per-cell SSB measurement duration $T_{\text{SS,0}}$ (for example, from a real 28 GHz prototype or a standard numerology) and check whether $K = 4$ still gives $\beta_{\text{sleep}} > 85\%$ at both frequencies; if $T_{\text{SS,0}}$ is a large fraction of the 20 ms SSB period, the $K = 4$ optimum shifts or disappears.

Watch

Extended reading notes

Core claim

The paper claims that, for a stationary UE under human and vehicular blockers modeled by double knife-edge diffraction, a listening set of size $K = 4$ out of nine cells is sufficient: the blocking probability $P_B$ drops below 1% at both 28 GHz and 140 GHz, and the fractional sleep time $\beta_{\text{sleep}}$ is maximized, exceeding 85% at both frequencies. This is the result of a simulation in which the UE measures synchronization bursts from the chosen cells each 20 ms DRX cycle, switches its serving cell to the best measured alternative when the current one falls below a threshold, and performs an exhaustive beam sweep only when all monitored links are blocked. The authors infer that $K = 4$ balances the cost of longer awake periods against the cost of frequent beam sweeps, and that macro-diversity with a small listening set can maintain reliable connectivity even where blockage dynamics are faster at 140 GHz.

Load-bearing premise

Equation (4) assumes the awake time to monitor $K$ cells is exactly $T_{\text{SS}} = K T_{\text{SS,0}}$ with a per-cell measurement time $T_{\text{SS,0}}$ that the paper never assigns a value, so the plotted sleep fractions and the more-than-85% claim rest on an unstated number.

Editorial extensions

If this is right

If $K = 4$ suffices at 140 GHz, a THz UE does not need to monitor all visible cells; a small listening set with occasional exhaustive beam sweep can meet a 99% availability target.
The greater-than-85% sleep fraction means the roughly 1 W front-end power at 140 GHz is drawn only about 15% of the time, substantially changing the energy budget for THz handhelds.
The proposed handoff rule lets a UE switch to the best monitored alternative before its current link fails, avoiding blind beam sweeps in most cycles.
The same optimal $K = 4$ emerging at both 28 and 140 GHz suggests that faster THz blockage can be countered by cell diversity rather than by monitoring more cells more often.

Reading between the lines

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

The model ignores beam-acquisition latency and array geometry; if measuring a cell actually takes longer than the assumed per-cell time $T_{\text{SS,0}}$, the optimal $K$ could shift above four in a hardware-realistic setting.
A natural extension would couple listening-set selection with bursty traffic scheduling, since the power saved by DRX also depends on how often the UE must wake to receive data.
The same listening-set idea could apply to multi-panel UE designs, where each panel points in a different direction and the panels themselves form the set to be scheduled for monitoring.
A testable prediction is that the $K = 4$ optimum will persist for non-stationary UEs only if the blocker velocity distribution and cell layout stay close to the simulated values; faster vehicle traffic should push the optimum upward.

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

First: this is a short, honest simulation study, and its qualitative takeaway—that monitoring a handful of cells at THz frequencies gets you most of the macro-diversity availability gain—is fine. The quantitative claim that K=4 gives >99% usable-link probability at both 28 and 140 GHz comes directly from the simulated blockage processes and does not depend on the timing model. That part survived my reading.

What is genuinely useful: it is one of the first attempts I know to put DRX power numbers and blockage-aware listening-set selection in the same frame at 140 GHz. The RFFE/ADC power comparison in Table I is a useful back-of-envelope, the 3GPP blockage model is applied carefully, and Algorithm 1 is simple enough to explain in one paragraph. The authors are also upfront that they extrapolate the 3GPP channel model past its stated 100 GHz range and that they assume perfect beam pointing via eigenbeamforming. The citation pattern looks clean—prior DRX work, 3GPP models, and their own prior power-model paper appear as inputs, not restatements.

The soft spot is the sleep-time claim. Equation (4) writes beta_awake >= (1-PB)*K*TSS0/TSS_per + PB, but TSS0, the time to measure one cell, never appears in Table II or anywhere in the text. Figure 6 is therefore not reproducible, and the position of the K=4 optimum is an artifact of a parameter the reader cannot check. The stress-test note is correct: with PB near 1%, the 85% sleep threshold is crossed only if TSS0 is about 0.73 ms; at 1 ms you get roughly 80% sleep, and at 5 ms the whole curve collapses. As TSS0 approaches zero, the optimum shifts to K=9. So the "saving the most amount of power" claim is not a finding—it is a hidden-input artifact. Also, the abstract and conclusion mention link prediction, but Algorithm 1 does no prediction; it just remeasures the same cells. That is a smaller, though real, overclaim.

This paper deserves a serious referee. The reliability result is meaningful, the problem is well placed in the literature, and the missing parameter is fixable. But a referee should require the value of TSS0, a sensitivity sweep over it, and ideally more than 100 trajectories before the power-saving headline can be accepted. Without that, the headline should be deleted and replaced by a conditional statement.

I would bring it to my reading group as a quick discussion of what makes a simulation claim reproducible, but I would not cite it until the timing parameter is supplied. Send it to peer review, with a demanding referee.

Referee Report

4 major / 4 minor

Summary. The paper studies connected-mode discontinuous reception (DRX) for mmWave and THz cellular systems, where the UE monitors a subset of cells during each DRX cycle. The authors first estimate receiver front-end power consumption at 28 and 140 GHz, then propose a heuristic algorithm that selects a listening set of K cells and switches serving cells based on measured SNR. They evaluate the algorithm with 3GPP-based end-to-end channel simulations that include human and vehicular blockers, and report the blocking probability PB and fractional sleep time βsleep as functions of K. The main claim is that K = 4 is sufficient to guarantee a usable link at least 99% of the time while maximizing sleep fraction (over 85%) at both 28 and 140 GHz.

Significance. If the results hold, the paper provides a useful preliminary demonstration that aggressive DRX can be compatible with blockage-prone directional links, which is an important question for power-constrained UE design in mmWave and THz systems. The power-consumption comparison between 28 and 140 GHz front ends is concrete, and the simulation setup is carefully anchored to 3GPP-style blockage and path-loss models. The proposed algorithm is simple and clearly specified, and the paper is honest in calling its estimates preliminary. The central quantitative claim, however, depends on at least one unstated parameter and on several idealizations that need to be reconciled before the results can be used as published.

major comments (4)

[Section III, Eq. (4) and Section V-B, Fig. 6] The headline sleep-time claim is not reproducible because the per-cell measurement time TSS,0 used in Eq. (4) is never specified. Table II lists many timing parameters but omits TSS,0, and the text only states that TSS = K·TSS,0. Figure 6's >85% sleep result at K=4 requires, with PB≈1% from Fig. 5 and TSS,per=20 ms, that TSS,0 be approximately 0.7 ms or less; for TSS,0 = 1 ms the K=4 sleep fraction falls to roughly 80%, and for larger TSS,0 the claimed optimum shifts or disappears. The authors must state the value of TSS,0 used to generate Fig. 6 and provide a sensitivity analysis showing how βsleep and the optimum K depend on it.
[Section V-A, footnote 1 and Section V-B] The 3GPP TR 38.901 channel and blockage models are validated only up to 100 GHz, yet the paper applies them at 140 GHz to produce the central 140 GHz results, including the K=4 conclusion. This extrapolation is acknowledged in a footnote, but the magnitude of the resulting uncertainty is not assessed. Since the paper's contribution is specifically the mmWave-versus-THz comparison, the quantitative PB and βsleep curves at 140 GHz rest on an unvalidated model; the authors should either justify the extrapolation with additional measurements or model comparisons, or clearly delimit the 140 GHz claims as illustrative.
[Section V-A, idealizations, and Section V-B, Figs. 5 and 6] The simulation assumes eigen-beamforming with perfect alignment between the UE and every BS (Section V-A), which removes from the model the beam-sweeping and beam-management overhead that is a central source of wake-up time and power consumption in directional DRX. Because the paper's core trade-off is between awake time and reliability, this idealization directly affects the quantitative sleep-time claim. The authors should discuss how imperfect beam alignment or periodic beam refinement would change TSS,0 and the resulting K=4 sleep fraction, or explicitly bound the effect.
[Section V-B, Figs. 5 and 6] The results are based on only 100 channel trajectories per frequency, and the figures do not show error bars or confidence intervals. The statement that K=4 is sufficient to meet the 99% reliability target is a point estimate from a small Monte Carlo sample; with 100 trajectories, a 1% blocking probability corresponds to roughly one blocked trajectory, so the estimate is noisy. The authors should report confidence intervals or use a larger number of trajectories to support the K=4 conclusion.

minor comments (4)

[Section III, text near Fig. 4] The sentence describing TSS reads 'an interval of durationTSS' with a missing space; please correct the typo.
[Section III, text after Eq. (5)] The phrase 'trade-off between a the listening set size' contains a stray article; it should read 'between the listening set size'.
[Section IV, Algorithm 1] The notation An is used both for the listening set and for the set of monitored cells; this is understandable but would benefit from a brief clarification in the pseudocode caption.
[Section II, Table I] The table caption says 'all units in mW', but the RFFE row in the table is given in dBm in Fig. 3; please ensure units are consistent across the table and figure.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the K=4 reliability/sleep result is a simulated output, not a restatement of inputs; only peripheral RFFE assumptions are self-cited.

full rationale

The paper's central derivation is Eq. (4), beta_awake >= (1 - PB) * K * TSS,0 / TSS,per + PB, combined with simulated blockage probabilities PB from 3GPP-based trajectories. The K=4 reliability claim comes directly from the simulated PB curve in Fig. 5, and the K=4 sleep optimum is computed from Eq. (4) rather than reverse-engineered from it. No equation redefines its target, and no fitted constant is relabeled as a prediction. The paper does cite its own prior work: [11] supplies the passive-circuit RFFE assumptions behind Table I, and [4], [6] support blockage dynamics. However, these self-citations are inputs to the power-consumption motivation and do not enter Eqs. (4)-(5), so they do not load-bear the DRX reliability or sleep-time claim. The main numerical weakness is reproducibility, not circularity: TSS,0, the per-cell measurement time that linearly controls Fig. 6, is never assigned a value, so the >85% sleep numbers and the location of the K=4 optimum cannot be checked from the manuscript as written. That is an omitted-parameter gap for correctness review, not a circular reduction of a prediction to its inputs. Score 2 reflects the peripheral self-citations and the unstated timing scale, while the central derivation remains independent.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper's central claim rests on standard channel and blockage models, but several parameters are chosen rather than derived, and one critical parameter, TSS,0, is absent entirely. No new physical entities are introduced.

free parameters (4)

Detection SNR threshold gamma_min = -6.5 dB
Chosen for the blocking probability and sleep fraction computations; the K=4 conclusion is conditional on this threshold.
Per-cell measurement time TSS,0 = not specified
Defines TSS = K*TSS,0 in Section III and controls the sleep percentages in Fig. 6, but no numerical value is given anywhere.
Blocker density lambda_b = 0.01 m^-2
Chosen from reference [15]; the blocker density controls outage frequency and the K=4 result is sensitive to it.
Blocker speed ranges for humans and vehicles = 0 to 1 m/s and 0 to 28 m/s
Taken from 3GPP recommendations but chosen without sensitivity analysis; faster or slower blockers would change the blockage dynamics and the required K.

assumptions (5)

domain assumption 3GPP TR 38.901 channel and blockage models are valid at 140 GHz above the standardized 100 GHz limit.
Used for all 140 GHz channel trajectories; the paper flags this extrapolation in a footnote and does not justify it physically.
domain assumption Perfect instantaneous eigenbeamforming between UE and BS.
Section V.A sets beamforming vectors to singular vectors, removing array geometry, beam acquisition latency, and pointing error from the SNR and sleep-time calculations.
domain assumption Link outage events are independent across cells in the listening set.
Eq. (3) multiplies per-cell outage probabilities; spatial correlations due to shared blockers are not modeled.
domain assumption When all monitored links fail, the UE remains awake for at least one entire DRX cycle to perform beam sweep.
Forms the beta_awake bound in Eq. (4), with an unquantified number L >= 1 of SSB periods for the sweep.
domain assumption The UE can measure K cells sequentially in K*TSS,0 time and sleep for the remainder of a 20 ms SSB period.
Sets the DRX timeline in Section III and determines the sleep fraction, but TSS,0 is never quantified.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Power Efficient Discontinuous Reception in THz and mmWave Wireless Systems." pith.science (2026). https://pith.science/paper/Z7KGPDKI

@misc{pith2026190808966,
  author       = {Pith},
  title        = {Pith review of: Power Efficient Discontinuous Reception in THz and mmWave Wireless Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Z7KGPDKI}},
  note         = {Machine review of arXiv:1908.08966}
}

read the original abstract

Discontinuous reception (DRX), where a user equip-ment (UE) temporarily disables its receiver, is a critical power saving feature in modern cellular systems. DRX is likely tobe particularly aggressively used in the mmWave and THzfrequencies due to the high front end power consumption. A keychallenge of DRX in these frequencies is that individual links are directional and highly susceptible to blockage. MmWave and THz UEs will therefore likely need to monitor multiple cells in multiple directions to ensure continuous reliable connectivity.This work proposes a novel, heuristic algorithm to dynamically select the cells to monitor to attempt to optimally trade-off link reliability and power consumption. The paper provides preliminary estimates of connected mode DRX mode consumption using detailed and realistic statistical models of blockers at both 28 and 140 GHz. It is found that although blockage dynamics are faster at 140 GHz, reliable connectivity at low power can be maintained with sufficient macro-diversity and link prediction

Figures

Figures reproduced from arXiv: 1908.08966 by the authors.

**Figure 2.** Analog beamforming based multi-channel receiver front end with [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 4.** Time line of a UE in connected mode DRX. [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: Variation of PB as a function of K. The dashed horizontal line corresponds to PB = 1%. 1 2 3 4 5 6 7 8 9 K 60 70 80 90 sleep (%) 28 GHz 140 GHz [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

**Figure 6.** Figure 6: Fractional UE sleep time, βsleep, as a function of K [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 13 canonical work pages

[1]

LTE/LTE-A discontinuous reception modeling for machine type communications,

J. Zhou, N. Nikaein, and T. Spyropoulos, “LTE/LTE-A discontinuous reception modeling for machine type communications,” IEEE Wireless Commun. Lett., vol. 2, no. 1, pp. 102-105 , 2013

work page 2013
[2]

Performance evaluation of LTE/LTE-a DRX: A Marko- vian approach,

H. Ramazanali, “Performance evaluation of LTE/LTE-a DRX: A Marko- vian approach,”IEEE Internet Things J., vol. 3, no. 3, pp. 386-397, 2016

work page 2016
[3]

T. S. Rappaport, R. W. Heath Jr, R. C. Daniels, and J. N. Murdock, Millimeter wave wireless communications . Pearson Education, 2014

2014
[4]

Empirical effects of dynamic human-body blockage in 60 GHz com- munications,

C. Slezak, V . Semkin, S. Andreev, Y . Koucheryavy, and S. Rangan, “Empirical effects of dynamic human-body blockage in 60 GHz com- munications,” IEEE Commun. Mag. , vol. 56, no. 12, pp. 60–66, 2018

work page 2018
[5]

Spatio-temporal impact of hand and body blockage for millimeter-wave user equipment design at 28 GHz,

V . Raghavan, V . Podshivalov, J. Hulten, M. A. Tassoudji, A. Sampath, O. H. Koymen, and J. Li, “Spatio-temporal impact of hand and body blockage for millimeter-wave user equipment design at 28 GHz,” IEEE Commun. Mag., vol. 56, no. 12, pp. 46–52, 2018

work page 2018
[6]

Rapid fading due to human blockage in pedestrian crowds at 5g millimeter-wave frequencies,

G. R. MacCartney, T. S. Rappaport, and S. Rangan, “Rapid fading due to human blockage in pedestrian crowds at 5g millimeter-wave frequencies,” in Proc. IEEE GLOBECOM , 2017, pp. 1–7

work page 2017
[7]

On the macro diversity with multiple bss to mitigate blockage in millimeter-wave communications,

J. Choi, “On the macro diversity with multiple bss to mitigate blockage in millimeter-wave communications,” IEEE Commun. Lett. , vol. 18, no. 9, pp. 1653–1656, 2014

work page 2014
[8]

TR 38.901, study on channel model for frequencies from 0.5 to 100 GHz (release 15) document,

3GPP, “TR 38.901, study on channel model for frequencies from 0.5 to 100 GHz (release 15) document,” Jun. 2018

work page 2018

Show all 15 references

[9]

An RF receiver for intra-band carrier aggre- gation,

S. Hwu and B. Razavi, “An RF receiver for intra-band carrier aggre- gation,” IEEE J. Solid-State Circuits , vol. 50, no. 4, pp. 946–961, Apr. 2015

2015
[10]

FCC online table of frequency allocations,

“FCC online table of frequency allocations,” Federal Communications Commission, Tech. Rep. 47 C.F.R, 2.106, 2018. [Online]. Available: https://transition.fcc.gov/oet/spectrum/table/fcctable.pdf

2018
[11]

A case for digital beamforming at mmWave,

S. Dutta, C. Barati, A. Dhananjay, D. A. Ramirez, J. F. Buckwalter, and S. Rangan, “A case for digital beamforming at mmWave,”arXiv preprint arXiv:1901.08693, 2019

1901 arXiv
[12]

TS 38.300, NR and NG-RAN Overall Description; Stage 2,

3GPP, “TS 38.300, NR and NG-RAN Overall Description; Stage 2,” 2017

2017
[13]

TS 38.331 NR - Radio Resource Control (RRC) protocol speciﬁcation - Release 15,

——, “TS 38.331 NR - Radio Resource Control (RRC) protocol speciﬁcation - Release 15,” 2017

2017
[14]

A tutorial on beam management for 3gpp nr at mmwave frequencies,

M. Giordani, M. Polese, A. Roy, D. Castor, and M. Zorzi, “A tutorial on beam management for 3gpp nr at mmwave frequencies,” arXiv preprint arXiv:1804.01908, 2018

2018 arXiv
[15]

Limited by capacity or blockage? a millimeter wave blockage analysis,

I. K. Jain, R. Kumar, and S. Panwar, “Limited by capacity or blockage? a millimeter wave blockage analysis,” Proc. Intl. Teletrafﬁc Congress , pp. 153–159, 2018

2018

Pith tools

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

[1] [1]

LTE/LTE-A discontinuous reception modeling for machine type communications,

J. Zhou, N. Nikaein, and T. Spyropoulos, “LTE/LTE-A discontinuous reception modeling for machine type communications,” IEEE Wireless Commun. Lett., vol. 2, no. 1, pp. 102-105 , 2013

work page 2013

[2] [2]

Performance evaluation of LTE/LTE-a DRX: A Marko- vian approach,

H. Ramazanali, “Performance evaluation of LTE/LTE-a DRX: A Marko- vian approach,”IEEE Internet Things J., vol. 3, no. 3, pp. 386-397, 2016

work page 2016

[3] [3]

T. S. Rappaport, R. W. Heath Jr, R. C. Daniels, and J. N. Murdock, Millimeter wave wireless communications . Pearson Education, 2014

2014

[4] [4]

Empirical effects of dynamic human-body blockage in 60 GHz com- munications,

C. Slezak, V . Semkin, S. Andreev, Y . Koucheryavy, and S. Rangan, “Empirical effects of dynamic human-body blockage in 60 GHz com- munications,” IEEE Commun. Mag. , vol. 56, no. 12, pp. 60–66, 2018

work page 2018

[5] [5]

Spatio-temporal impact of hand and body blockage for millimeter-wave user equipment design at 28 GHz,

V . Raghavan, V . Podshivalov, J. Hulten, M. A. Tassoudji, A. Sampath, O. H. Koymen, and J. Li, “Spatio-temporal impact of hand and body blockage for millimeter-wave user equipment design at 28 GHz,” IEEE Commun. Mag., vol. 56, no. 12, pp. 46–52, 2018

work page 2018

[6] [6]

Rapid fading due to human blockage in pedestrian crowds at 5g millimeter-wave frequencies,

G. R. MacCartney, T. S. Rappaport, and S. Rangan, “Rapid fading due to human blockage in pedestrian crowds at 5g millimeter-wave frequencies,” in Proc. IEEE GLOBECOM , 2017, pp. 1–7

work page 2017

[7] [7]

On the macro diversity with multiple bss to mitigate blockage in millimeter-wave communications,

J. Choi, “On the macro diversity with multiple bss to mitigate blockage in millimeter-wave communications,” IEEE Commun. Lett. , vol. 18, no. 9, pp. 1653–1656, 2014

work page 2014

[8] [8]

TR 38.901, study on channel model for frequencies from 0.5 to 100 GHz (release 15) document,

3GPP, “TR 38.901, study on channel model for frequencies from 0.5 to 100 GHz (release 15) document,” Jun. 2018

work page 2018