Pith. sign in

REVIEW 3 major objections 5 minor 40 references

Viveka: Context-Aware Sensing for Energy Efficiency in Smart Wearables

T0 review · 3 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read Viveka claims a context-gated sensing policy can cut wearable energy use by up to 75% while keeping activity-recognition accuracy within a few percent of always-on baselines.

desk verdict Sensible gating design, but the headline savings are best-case; needs nested cross-validation and a cleaner NP-hardness proof before the numbers can be trusted. read the letter →

arxiv 2608.05572 v1 pith:G7YHLRTF submitted 2026-08-06 cs.ET cs.DC

Nikhil Sreekumar , Abhishek Chandra This is my paper · ORCID
classification cs.ETcs.DC
keywords wearablesensingenergyefficiencycontext-awaresensorselectionadaptivesamplinghumanactivityrecognitionpermutationfeatureimportancespectralanalysis
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 tries to show that joint sensor selection and sampling-rate adaptation can be made both safe and practical for wearables by separating context estimation from trust estimation. A light, always-on controller decides not just what activity is happening but how much to trust that guess, and a three-tier gating policy applies aggressive per-context sensor and rate settings only when confidence is high. This turns a risky optimization into a closed-loop decision problem, and the reported evaluations on MHEALTH and PAMAP2 indicate large energy and data savings with only minor accuracy loss. If correct, the framework addresses a practical blocker for smart wearables: battery life and the need to avoid acting on a misidentified context.

What carries the argument

The central mechanism is the stability-and-confidence gate: a three-tier decision rule that applies an aggressive per-context sensor and sampling-rate configuration only when the lightweight controller's recent activity predictions are both repeated and above a confidence threshold, and otherwise widens the sensor set or falls back to a conservative default. This gate is supported by an always-on controller that estimates both context and its own trustworthiness, offline policy maps built by permutation feature importance for sensor sets and FFT-based spectral energy analysis for per-sensor rates, a rate-degradation augmented training scheme that matches online reconstruction, and a datasheet-grounded energy model spanning sensing, computation, and BLE communication.

What would settle it

Deploy the full policy on a real wearable instrumented to measure per-sensor current draw and total battery life while a subject follows the MHEALTH and PAMAP2 protocols, and check whether measured energy savings and F1 fall within the reported envelope (up to 75% energy, 78% data reduction, accuracy within 3-5%); separately, enumerate all configurations of a small APSSE instance with a real trained classifier to test whether the minimal-energy configuration matches the coverage equivalence used in the NP-hardness reduction.

Watch

Extended reading notes

Core claim

Viveka's central claim is that applying an aggressive, context-specific sensing policy is safe when the on-device context estimate is both stable and confident. The system runs a lightweight controller that continuously outputs an activity prediction and a confidence value; a stability-and-confidence gate then chooses one of three tiers: the fully optimized per-activity configuration when recent predictions are consistent and confident, a recovery configuration that unions the sensor requirements of plausible activities under moderate uncertainty, and a conservative high-fidelity fallback when context is unresolved. Per-activity configurations are computed offline using permutation feature importance to choose minimal sensor sets and spectral (Nyquist-based) energy analysis to choose per-sensor sampling rates, and the recognition model is trained with rate-degraded, reconstructed-signal augmentation so it stays accurate on the subsampled input it receives at runtime. On MHEALTH and PAMAP2, the paper reports up to 75% energy savings and 78% data reduction over standard baselines while keeping F1 within 3-5% of the always-on baseline, with projected battery life extending from 8.7 to 34.6 days on a 300 mAh battery for MHEALTH.

Load-bearing premise

The NP-hardness proof assumes that the classifier error function can be defined so that meeting the accuracy constraint is equivalent to covering a set of activity types with the active sensors, but no concrete classifier, data distribution, or accuracy model is supplied to realize that error function.

Editorial extensions

If this is right

  • If the reported results hold, wearable devices can extend operational battery life substantially, with the paper's projections roughly quadrupling runtime on MHEALTH under a 300 mAh battery.
  • Data volume transmitted to the edge drops by roughly three quarters, reducing BLE energy use and on-device storage pressure while still feeding the recognition model reconstructed signals.
  • The gating idea generalizes to other energy-constrained multi-sensor IoT settings where acting on a misidentified context is risky, since the controller is decoupled from the recognition model.
  • Because joint sensor and rate selection is NP-hard, a runtime-tractable heuristic like Viveka is the practical route; future exact or approximate solvers would need a different basis.
  • Training the recognition model on rate-degraded, reconstructed signals closes the gap between offline policy computation and online execution, making sub-sampled sensing viable for deployment.
  • The three-tier fallback behavior avoids sensor thrashing, addressing the energy overhead of rapidly toggling sensors on and off during activity transitions.

Reading between the lines

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

  • Editorial inference: the headline 75% energy and 78% data figures reflect a best-case threshold configuration selected on a single subject-disjoint split, while the paper's leave-one-subject-out numbers are more conservative (approximately 64% on MHEALTH and 49% on PAMAP2), so the practical savings depend on how robustly the thresholds are chosen without tuning on test subjects.
  • Editorial inference: if the confidence gate works as described, a natural extension is to make the threshold itself adaptive to battery level or user history, trading a little accuracy for longer operation when the battery is low — a testable design the paper does not pursue.
  • Editorial inference: the offline per-activity sensor importance assumes the activity-conditional feature ranking is stable across users and deployment conditions; personalized or online-updated importance maps could either improve savings or expose sensitivity to distribution shift.
  • Editorial inference: the NP-hardness reduction uses a coverage-based placeholder for the classifier error function; the heuristic's value does not depend on this proof, but a more realistic error model could change the problem's complexity and should be examined separately.
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

3 major / 5 minor

Summary. The paper formulates joint sensor selection and sampling-rate selection as an optimization problem (APSSE) under a classification-error constraint, claims it is NP-hard, and proposes Viveka, a lightweight context-aware framework. Viveka uses a low-power controller to estimate activity and confidence, a stability/confidence-gated policy with three tiers (optimized per-context policy, recovery union policy, conservative fallback), and offline per-context sensor/rate policies derived from permutation feature importance and FFT-based spectral analysis. The framework is evaluated on the MHEALTH and PAMAP2 datasets, with reported best-case energy savings up to 75% and data reduction up to 78% while keeping F1 within 3-5% of baselines.

Significance. If the reported results hold under a properly validated protocol, the paper presents a practically relevant and well-motivated approach to energy-efficient wearable sensing. The tiered gating idea is a genuine contribution over static selection and duty-cycling baselines: it directly addresses the risk of applying an aggressive sensing policy to a misidentified context. The paper also provides a concrete, datasheet-grounded energy model, a sensitivity analysis over the four thresholds and decision interval, and a designed co-training step with rate-degraded augmentation. These are strengths that go beyond a purely algorithmic contribution. However, the central quantitative claim currently rests on a best-case configuration whose thresholds were selected on a single subject-disjoint split, as disclosed in the Section I footnote, and the paper reports no variance or per-subject statistics. The contribution is therefore promising but not yet supported at the level claimed in the abstract and conclusion.

major comments (3)
  1. [§I (footnote 2), §V-D, §V-E, Fig. 7, Table III] The headline numbers (75% energy savings and 78% data reduction in the abstract and conclusion) are attributed to a 'best-case configuration' in which policy thresholds were selected on a single subject-disjoint split (footnote 2 in Section I). Section V-E describes parameter sweeps but does not state that threshold selection was restricted to a held-out validation set; instead, Section V-D reports the resulting dataset-specific hyperparameters as though they are fixed. The footnote's own leave-one-subject-out numbers (≈64% on MHEALTH and ≈49% on PAMAP2) are materially lower than the headline, so the central empirical claim is not yet protected against test-set selection bias. This is load-bearing because the paper's main contribution is the claimed energy/accuracy trade-off. Please re-run the evaluation under a nested leave-one-subject-out protocol in which all four thresholds and T_interval are chosen only on training subjects, and report the resulting Figure 7, Table III, and F1 comparisons.
  2. [Appendix VIII, Theorem 1] The NP-hardness reduction constructs the classifier error function C(x) so that C(x) ≤ C_threshold holds exactly when the selected active sensors cover all elements of U. This is a by-construction equivalence, not a derivation from a concrete classifier, data distribution, and accuracy model. Since C(x) in the original formulation (Eq. (5)) is the expected classification error of the inference model, the reduction should show that a coverage-indicator error function is realizable in that setting; otherwise the proof only establishes hardness for a specially constructed variant of APSSE. This affects the stated contribution 'we prove it is NP-hard.' Please either instantiate the reduction with a concrete classifier and data distribution, or explicitly restate the theorem for a class of problems with arbitrary error functions C(x).
  3. [§V-D, §V-E, Table III] No variance, confidence interval, or per-subject statistic is reported for any energy, data-reduction, or F1 metric. With only 10 (MHEALTH) and 9 (PAMAP2) subjects, the reported differences between Viveka and baselines, e.g., F1 0.758 vs 0.683 and energy savings 74% vs 72% against GS-VS on MHEALTH, may be within subject-level noise. The projected battery-life numbers in Table III are deterministic point estimates. A per-subject breakdown or confidence intervals are needed to support the comparative claims and to let readers assess the stability of the savings across users.
minor comments (5)
  1. [Abstract, Section VII] The abstract and conclusion repeat the best-case savings figures without the caveat disclosed in the Section I footnote; please either add the qualification in these locations or report the leave-one-subject-out numbers as the primary results.
  2. [Section V-G, Figure 10] The claim that 'Transient periods are captured with distinct sensor configurations' is qualitative; please provide a quantitative measure of transition detection latency or the fraction of transitions correctly detected.
  3. [Section II-B, Figures 3 and 4] The WISDM PFI experiment reports accuracy drops without error bars or repeated-run variance; adding this would increase confidence in the claim that sensor importance is activity-dependent.
  4. [Section VI, Reference [18]] The CoSS reference appears incomplete: no venue or publication year is given.
  5. [Throughout] There are several typographical and grammatical errors, e.g., 'The nameVivekais' (missing space), 'uncertainity', and 'upto'; a careful proofread is needed.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline savings are measured with thresholds tuned on the test split; the paper's own footnote shows leave-one-subject-out savings are materially lower.

  1. fitted input called prediction [Section I footnote; Section V-D 'Overall Impact of Viveka'; Section V-E 'Parameter Tuning']
    "These figures reflect a best-case configuration in which policy thresholds were selected on a single subject-disjoint split. Subsequent leave-one-subject-out cross-validation, with thresholds selected only on held-out training subjects, yields more conservative savings (approximately 64% on MHEALTH and 49% on PAMAP2) while preserving comparable accuracy; a full treatment is the subject of forthcoming work. ... The values of the parameters were decided based on the parameter tuning experiments discussed in the Section V-E."

    The paper uses a single subject-disjoint train/test split (Section V-A: 80% training, 20% testing by subject) and describes no validation split. Section V-E then selects each of the four policy thresholds by optimizing energy, data-reduction, and F1 metrics, and Section V-D reports the headline savings ('74%' energy / '73%' data on MHEALTH; '62%' / '78%' on PAMAP2) using exactly those selected thresholds. The reported numbers are therefore best-case selections scored on the same evaluation split, not out-of-sample predictions: the thresholds that maximize the metrics are the ones being evaluated.

full rationale

Viveka's core mechanism is not a renamed derivation: sensor selection via permutation feature importance, rate selection via spectral energy analysis, and the gating policy are concrete, externally checkable procedures, and the energy model uses commercial datasheets plus measured inference costs. The evaluation against AlwaysOn, VS, GS, GS-VS, AS, AR, and AS-VS is an empirical measurement rather than a formal derivation. I do not count the Appendix VIII NP-hardness proof as circular: a reduction may legitimately define an oracle error function, although the proof leaves open whether such a C(x) is realizable by an actual classifier—that is a rigor gap, not a circularity. The load-bearing circularity is in the headline evaluation: the thresholds used to produce the 75%/78% figures were tuned on the same test split that reports those figures, making them selected best cases rather than predictive results. The paper's own footnote quantifies the gap (LOSO yields roughly 64%/49%), confirming that the headline numbers are partly an artifact of test-set tuning. Because the paper discloses this and the conservative numbers remain substantial, the circularity is partial rather than total.

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

The central engineering claims rest on the thresholds listed below, which are tuned per dataset; the NP-hardness contribution rests on an ad hoc construction of the classifier error function. The energy model and the calibration of the confidence scores are domain assumptions. No new physical entities are introduced.

free parameters (5)
  • PFI threshold (theta_PFI) = 0.2 (MHEALTH), 0.05 (PAMAP2)
    Threshold for permutation feature importance to decide which sensors to keep per activity; tuned on the evaluated datasets (Section V-E).
  • FFT energy threshold (theta_FFT) = 0.95 (MHEALTH), 0.8 (PAMAP2)
    Energy threshold for spectral analysis to set per-sensor sampling rates; tuned on the evaluated datasets (Section V-E).
  • Confidence threshold (theta_conf) = 0.7 (both datasets)
    Threshold for applying the optimized per-context policy during stable, confident periods; tuned on the evaluated datasets (Section V-E).
  • Fallback confidence threshold (beta_conf) = 0.1 (MHEALTH), 0.0 (PAMAP2)
    Threshold for selecting candidate activities in the intermediate fallback policy; tuned on the evaluated datasets (Section V-E).
  • Decision interval (T_interval) = 2 seconds (both datasets)
    Interval at which the wearable model re-evaluates the context; chosen from a sweep over [1,2,4,8,16] seconds (Section V-F).
assumptions (4)
  • domain assumption The sensing energy, compute energy, and BLE transmission energy are additive and dominated by the components modeled in Equation (9).
    Section V-C defines E_total as the sum of three components and treats datasheet values as representative. If the additive model or constant power draw assumptions are wrong, the projected battery life and savings numbers change.
  • standard math Nyquist-Shannon sampling theorem is applicable to the per-activity sensor signals, so the rate computed from the FFT energy threshold suffices for the downstream classifier.
    Algorithm 2 uses FFT plus Nyquist to pick rates; this assumes the classifier only needs the band-limited signal content up to the energy threshold.
  • ad hoc to paper There exists a classifier error function C(x) for the APSSE problem that behaves as a set-cover indicator in the NP-hardness reduction.
    Appendix VIII, step 4 defines C(x) such that the constraint is satisfied iff sensors cover universe U, without giving a concrete model. This is an ad hoc construction that makes the reduction valid only under this unverified condition.
  • domain assumption The wearable model's softmax confidence scores are reliable enough for the fixed thresholds theta_conf and beta_conf to separate certain and uncertain contexts.
    Algorithm 3 uses raw softmax probabilities as confidence; no calibration step is described, so miscalibrated probabilities would change the gating behavior.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Viveka: Context-Aware Sensing for Energy Efficiency in Smart Wearables." pith.science (2026). https://pith.science/paper/G7YHLRTF

@misc{pith2026260805572,
  author       = {Pith},
  title        = {Pith review of: Viveka: Context-Aware Sensing for Energy Efficiency in Smart Wearables},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G7YHLRTF}},
  note         = {Machine review of arXiv:2608.05572}
}
read the original abstract

The proliferation of multi-sensor Internet of Things (IoT) systems, from Body Sensor Networks (BSNs) to industrial monitoring, is increasingly constrained by strict energy budgets and limited on-device storage. Continuous high-fidelity sensing leads to rapid battery depletion and data gaps that compromise application reliability. Existing strategies address this through sensor selection or adaptive sampling in isolation, or rely on computationally expensive agents for joint optimization. They lack context granularity or introduce significant overhead, and critically, they do not account for the risk that an aggressive, context-specific sensing policy applied to a misidentified context degrades accuracy. In this paper, we formulate joint sensor and sampling-rate selection as an NP-hard energy-minimization problem and propose Viveka, a lightweight, context-aware framework. Viveka couples a cheap, always-on controller that estimates context and how much to trust that estimate with a stability and confidence gated policy that applies an aggressive per-context configuration only when context is certain, and falls back safely otherwise. Per-context configurations are instantiated using permutation feature importance and spectral energy analysis. Evaluation on the MHEALTH and PAMAP2 datasets shows that Viveka achieves up to 75% energy savings and 78% data reduction over standard baselines in a best-case configuration, while maintaining classification accuracy within 3-5% of the baselines.

Figures

Figures reproduced from arXiv: 2608.05572 by the authors.

Figure 1
Figure 1. Battery of commercial smart wearables. pace with the growth of sensor capabilities and processing demands [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 3
Figure 3. PFI for the WISDM dataset at 20Hz acc\_x acc\_y acc\_z gyro\_x gyro\_y gyro\_z Feature A B C D E F G H I J K L M O P Q R S Activity 0.2720 0.2462 0.0780 0.4159 0.4068 0.6795 0.2879 0.0674 0.0432 0.0015 0.3591 0.5462 0.1348 0.0689 0.1205 0.2348 0.1833 0.2341 0.1925 0.3185 0.2891 0.0906 0.0551 0.0604 0.1306 0.1660 0.2189 0.0211 0.0415 0.0513 0.0652 0.1508 0.0598 0.0356 0.0538 0.0174 0.0581 0.0747 -0.0030 0.0823 0.0249… view at source ↗
Figure 4
Figure 4. PFI for the WISDM dataset at 10Hz The sensor sampling rate also provides an opportunity for energy savings, as it can often be reduced without a major impact on performance. To demonstrate this, we extended our previous experiment using the same LSTM model trained on the 20 Hz WISDM dataset. During testing, we downsampled the data to mimic 10 Hz sensing by selecting every other data point. We then interpolated this … view at source ↗
Figures from the paper (7 more)
Figure 5
Figure 5. Figure 5: System Model A. System Components The system we consider consists of multiple sensors that reside on a single edge device, such as a smartwatch, ring, or band. Alternatively, the system could be a body sensor network (BSN) where sensors connect to a central gateway dev…
Figure 6
Figure 6. Figure 6: Viveka framework IV. VIVEKA In this section, we propose Viveka, a novel, context-aware framework that selects sensors and corresponding sampling rates based on the context to reduce the energy consumption associated with sensing. Since the APSSE joint optimization is N…
Figure 7
Figure 7. Figure 7: Viveka achieves the highest energy and data reduction [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: 1) PFI Threshold: The PFI threshold (θP F I ) controls which sensors are retained in the per-activity policy, lower values retain more sensors. The θP F I is varied across nine discrete values ranging from 0.01 to 1.0 {0.01, 0.02, 0.05, 0.1, 0.2, 0.4, 0.6, 0.8, 1.0}. O…
Figure 8
Figure 8. Figure 8: Parameter Tuning. even on uncertain windows, occasionally feeding the classifier degraded or incomplete feature vectors and lowering F1. Too high a threshold keeps the system in its conservative fallback state, where more sensors remain active at default rates. θconf =…
Figure 9
Figure 9. Figure 9: Decision interval. As decision interval increases, energy [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Sensor activations for MHEALTH. The top ribbon displays the ground truth activity timeline, while the heatmap [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 36 canonical work pages

  1. [1]

    Wearables and internet of things (iot) technologies for fitness assessment: a systematic review,

    J. Passos, S. I. Lopes, F. M. Clemente, P. M. Moreira, M. Rico-Gonz ´alez, P. Bezerra, and L. P. Rodrigues, “Wearables and internet of things (iot) technologies for fitness assessment: a systematic review,”Sensors, vol. 21, no. 16, p. 5418, 2021

    work page 2021

  2. [2]

    The golden age of bio-logging: How animal-borne sensors are advancing the frontiers of ecology,

    C. C. Wilmers, B. Nickel, C. M. Bryce, J. A. Smith, R. E. Wheat, and V . Yovovich, “The golden age of bio-logging: How animal-borne sensors are advancing the frontiers of ecology,”Ecology, vol. 96, no. 7, pp. 1741–1753, 2015

    work page 2015

  3. [3]

    R. O’Dwyer. (2022) Shell trials crew smartwatches for safety management. Accessed on May 1, 2026. [Online]. Available: https://smartmaritimenetwork.com/2022/05/03/ shell-trials-crew-smartwatches-for-safety-management/

    work page 2022

  4. [4]

    Advancing a us navy shipboard infrastructure for sleep monitoring with wearable technology,

    A. G. Kubala, P. G. Roma, J. T. Jameson, P. H. Sessoms, E. D. Chinoy, L. R. Rosado, T. B. Viboch, B. J. Schrom, H. N. Rizeq, P. S. Gordyet al., “Advancing a us navy shipboard infrastructure for sleep monitoring with wearable technology,”Applied Ergonomics, vol. 117, p. 104225, 2024

    work page 2024

  5. [5]

    Wearables for industrial work safety: A survey,

    E. Svertoka, S. Saafi, A. Rusu-Casandra, R. Burget, I. Marghescu, J. Hosek, and A. Ometov, “Wearables for industrial work safety: A survey,”Sensors, vol. 21, no. 11, p. 3844, 2021

    work page 2021

  6. [6]

    Minimum-Cost Sensor Channel Selection For Wearable Computing

    R. K. Sah and H. Ghasemzadeh, “Minimum-cost sensor channel selec- tion for wearable computing,”arXiv preprint arXiv:2402.00875, 2024

    work page Pith review arXiv 2024

  7. [7]

    Sensor-classifier co-optimization for wearable human activity recognition applications,

    N. Anish, G. Bhat, J. Park, H. G. Lee, and U. Y . Ogras, “Sensor-classifier co-optimization for wearable human activity recognition applications,” in2019 IEEE International Conference on Embedded Software and Systems (ICESS). IEEE, 2019, pp. 1–4

    work page 2019

  8. [8]

    Effective sensor selection for human activity recognition via shapley value,

    E. Borella, U. B. C ¸ akmakc ¸i, E. Gottardis, A. Buratto, T. Marchioro, and L. Badia, “Effective sensor selection for human activity recognition via shapley value,” in2024 IEEE International Workshop on Metrology for Living Environment (MetroLivEnv). IEEE, 2024, pp. 22–27

    work page 2024

Show all 40 references
  1. [9]

    Optimizing sensor deployment for multi-sensor- based har system with improved glowworm swarm optimization algo- rithm,

    Y . Tian and J. Zhang, “Optimizing sensor deployment for multi-sensor- based har system with improved glowworm swarm optimization algo- rithm,”Sensors, vol. 20, no. 24, p. 7161, 2020

    2020

  2. [10]

    Vfds: variational foresight dynamic selection in bayesian neu- ral networks for efficient human activity recognition,

    R. Ardywibowo, S. Boluki, Z. Wang, B. J. Mortazavi, S. Huang, and X. Qian, “Vfds: variational foresight dynamic selection in bayesian neu- ral networks for efficient human activity recognition,” inInternational Conference on Artificial Intelligence and Statistics. PMLR, 2022, ...

    2022

  3. [11]

    Dana: Dimension-adaptive neural architecture for multivariate sensor data,

    M. Malekzadeh, R. Clegg, A. Cavallaro, and H. Haddadi, “Dana: Dimension-adaptive neural architecture for multivariate sensor data,” Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiq- uitous Technologies, vol. 5, no. 3, pp. 1–27, 2021

    2021

  4. [12]

    Adaptive sampling in context-aware systems: a machine learning approach,

    A. L. Wood, G. V . Merrett, S. R. Gunn, B. M. Al-Hashimi, N. R. Shadbolt, and W. Hall, “Adaptive sampling in context-aware systems: a machine learning approach,” inIET Conference on Wireless Sensor Systems (WSS 2012). IET, 2012, pp. 1–5

    2012

  5. [13]

    Freqsense: Adaptive sampling rates for sensor-based human activity recognition under tunable computational budgets,

    G. Yang, L. Zhang, C. Bu, S. Wang, H. Wu, and A. Song, “Freqsense: Adaptive sampling rates for sensor-based human activity recognition under tunable computational budgets,”IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 12, pp. 5791–5802, 2023

    2023

  6. [14]

    On adaptive sampling algorithms for iot devices,

    Y . Ben-Aboud, D. B. Licea, M. Ghogho, and A. Kobbane, “On adaptive sampling algorithms for iot devices,” inICC 2021-IEEE International Conference on Communications. IEEE, 2021, pp. 1–7

    2021

  7. [15]

    Energy-aware adaptive sampling for self-sustainability in resource-constrained iot devices,

    M. Giordano, S. Cortesi, P.-V . Mekikis, M. Crabolu, G. Bellusci, and M. Magno, “Energy-aware adaptive sampling for self-sustainability in resource-constrained iot devices,” inProceedings of the 11th Inter- national Workshop on Energy Harvesting & Energy-Neutral Sensing System...

    2023

  8. [16]

    An adaptive sensor data segments selection method for wearable health care services,

    S.-Y . Chen, C.-F. Lai, R.-H. Hwang, Y .-H. Lai, and M.-S. Wang, “An adaptive sensor data segments selection method for wearable health care services,”Journal of medical systems, vol. 39, no. 12, p. 194, 2015

    2015

  9. [17]

    Instance-wise dynamic sensor selection for human activity recognition,

    X. Yang, Y . Chen, H. Yu, Y . Zhang, W. Lu, and R. Sun, “Instance-wise dynamic sensor selection for human activity recognition,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 34, no. 01, 2020, pp. 1104–1111

    2020

  10. [18]

    Coss: Co-optimizing sensor and sampling rate for data-efficient ai in human activity recognition,

    M. Liu, Z. Zhao, D. Geißler, B. Zhou, S. Suh, and P. Lukowicz, “Coss: Co-optimizing sensor and sampling rate for data-efficient ai in human activity recognition,” in2nd Workshop on Sustainable AI

  11. [19]

    A survey on human activity recognition using wearable sensors,

    O. D. Lara and M. A. Labrador, “A survey on human activity recognition using wearable sensors,”IEEE communications surveys & tutorials, vol. 15, no. 3, pp. 1192–1209, 2012

    2012

  12. [20]

    Wearables for health and fitness: Measurement characteristics and accuracy,

    L. Scalise and G. Cosoli, “Wearables for health and fitness: Measurement characteristics and accuracy,” in2018 IEEE international instrumenta- tion and measurement technology conference (I2MTC). IEEE, 2018, pp. 1–6

    2018

  13. [21]

    (2024) Sensing solutions for wearables. Bosch. Accessed on May 1, 2026. [Online]. Available: https://www.bosch-sensortec.com/ applications-solutions/wearables/

    2024

  14. [22]

    WISDM Smartphone and Smartwatch Activity and Bio- metrics Dataset ,

    G. Weiss, “WISDM Smartphone and Smartwatch Activity and Bio- metrics Dataset ,” UCI Machine Learning Repository, 2019, DOI: https://doi.org/10.24432/C5HK59

    2019 doi

  15. [23]

    Random forests,

    L. Breiman, “Random forests,”Machine learning, vol. 45, no. 1, pp. 5–32, 2001

    2001

  16. [24]

    mhealthdroid: a novel framework for agile development of mobile health applications,

    O. Banos, R. Garcia, J. A. Holgado-Terriza, M. Damas, H. Pomares, I. Rojas, A. Saez, and C. Villalonga, “mhealthdroid: a novel framework for agile development of mobile health applications,” inInternational workshop on ambient assisted living. Springer, 2014, pp. 91–98

    2014

  17. [25]

    Design, implementation and validation of a novel open framework for agile development of mobile health applications,

    O. Banos, C. Villalonga, R. Garcia, A. Saez, M. Damas, J. Holgado- Terriza, S. Lee, H. Pomares, and I. Rojas, “Design, implementation and validation of a novel open framework for agile development of mobile health applications,”BioMedical Engineering OnLine, vol. 14, pp. 1–20, 2015

    2015

  18. [26]

    Introducing a new benchmarked dataset for activity monitoring,

    A. Reiss and D. Stricker, “Introducing a new benchmarked dataset for activity monitoring,” in2012 16th international symposium on wearable computers. IEEE, 2012, pp. 108–109

    2012

  19. [27]

    Tinierhar: Towards ultra-lightweight deep learning models for efficient human activity recognition on edge devices,

    S. Bian, M. Liu, V . F. Rey, D. Geissler, and P. Lukowicz, “Tinierhar: Towards ultra-lightweight deep learning models for efficient human activity recognition on edge devices,”arXiv preprint arXiv:2507.07949, 2025

    2025 arXiv

  20. [28]

    Adasense: Adapting sampling rates for activity recognition in body sensor networks,

    X. Qi, M. Keallyet al., “Adasense: Adapting sampling rates for activity recognition in body sensor networks,” in2013 IEEE 19th Real-Time and Embedded Technology and Applications Symposium (RTAS). IEEE, 2013, pp. 163–172

    2013

  21. [29]

    BMI160 Data sheet,

    Bosch Sensortec, “BMI160 Data sheet,” Tech. Rep., November 2020, document revision 1.0, Document number BST-BMI160-DS000-09, Accessed on May 1,

    2020

  22. [30]

    BMM350: 3-axis magnetic sensor with high data rate,

    Bosch Sensortec , “BMM350: 3-axis magnetic sensor with high data rate,” Tech. Rep., February 2025, document revision 1.27, Document number BST-BMM350-DS001-27, Accessed on May 1,

    2025

  23. [31]

    MAX30003 - Ultra-Low Power, Single-Channel Integrated Biopotential (ECG, R-to-R Detection) AFE,

    Analog Devices, “MAX30003 - Ultra-Low Power, Single-Channel Integrated Biopotential (ECG, R-to-R Detection) AFE,” Tech. Rep., January 2018, rev 0; 1/18, Accessed on May 1, 2026. [Online]. Available: https://www.analog.com/media/en/technical-documentation/ data-sheets/max30003.pdf

    2018

  24. [32]

    Available: https://www.bosch-sensortec.com/media/ boschsensortec/downloads/datasheets/bst-bmm350-ds001.pdf

    [Online]. Available: https://www.bosch-sensortec.com/media/ boschsensortec/downloads/datasheets/bst-bmm350-ds001.pdf

  25. [33]

    nRF52840 Product Specification,

    Nordic Semiconductor, “nRF52840 Product Specification,” accessed on May 1, 2026. [Online]. Available: https://docs.nordicsemi.com/bundle/ ps nrf52840/page/keyfeatures html5.html

    2026

  26. [34]

    NST112 High Accuracy, Low-Power Temperature Sensor with SMBus and I2C Interface,

    NOVOSENSE Microelectronics, “NST112 High Accuracy, Low-Power Temperature Sensor with SMBus and I2C Interface,” Tech. Rep., 2023, rev. 2.1, Accessed on May 1, 2026. [Online]. Available: https://www. novosns.com/files/NST112x-CWLR Datasheet Rev 2.1 EN.pdf

    2023

  27. [36]

    An adaptive algorithm to improve en- ergy efficiency in wearable activity recognition systems,

    H. Rezaie and M. Ghassemian, “An adaptive algorithm to improve en- ergy efficiency in wearable activity recognition systems,”IEEE Sensors Journal, vol. 17, no. 16, pp. 5315–5323, 2017. VIII. NP-HARDNESSPROOF Theorem 1.The APSSE problem formulated in equations(4)– (7)is NP-Hard...

    2017

  28. [37]

    The condition C(x)≤C threshold is equivalent to covering all necessary activity types (i.e., achieving 100% required accuracy implies covering all elements inU)

    Mapping Activities to Universe: Let the UniverseU represent the set of activity types required. The condition C(x)≤C threshold is equivalent to covering all necessary activity types (i.e., achieving 100% required accuracy implies covering all elements inU)

  29. [38]

    Mapping Sensors to Sets: Let each sensorsrepresent a setS i in the WSC problem

  30. [39]

    •Choosing indexj= 0(OFF) corresponds tonot selectingsetS i

    Mapping Sampling Rates to Selection: Restrict each sensorsto have only two rates:R s ={r s,0, rs,1}where rs,0 = 0andr s,1 =r high for simplicity. •Choosing indexj= 0(OFF) corresponds tonot selectingsetS i. Cost = 0. •Choosing indexj= 1(ON) corresponds toselecting setS i. Cost ...

  31. [40]

    If any elementu∈Uis uncovered, the error remains above the threshold

    Mapping Accuracy to Coverage: We define the error functionC(x)such thatC(x)≤C threshold is satisfied if and only ifthe union of features provided by the selected active sensors fully covers the universeU. If any elementu∈Uis uncovered, the error remains above the threshold. Co...

  32. [2026]

    Available: https://www.bosch-sensortec.com/media/ boschsensortec/downloads/datasheets/bst-bmi160-ds000.pdf

    [Online]. Available: https://www.bosch-sensortec.com/media/ boschsensortec/downloads/datasheets/bst-bmi160-ds000.pdf

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.