pith. sign in

arxiv: 2603.12278 · v2 · submitted 2026-02-27 · 🧬 q-bio.OT · cs.AI· cs.LG

Unsupervised Anomaly Detection in Wearable Foot Sensor Data: A Baseline Feasibility Study Towards Diabetic Foot Ulcer Prevention

Md Tanvir Hasan Turja This is my paper

Pith reviewed 2026-05-15 18:44 UTC · model grok-4.3

classification 🧬 q-bio.OT cs.AIcs.LG
keywords anomaly detectiondiabetic foot ulcerwearable sensorsunsupervised learningfoot temperatureplantar pressurebaseline modelfeasibility study
0
0 comments X

The pith

Unsupervised anomaly detection on foot temperature and pressure data from healthy adults creates a validated baseline pipeline for future diabetic foot ulcer studies.

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

The paper tests whether two unsupervised algorithms can flag statistical deviations in wearable sensor readings of foot temperature and plantar pressure collected from healthy people during hundreds of walking sessions. The goal is to define a reference model of normal foot biomechanics that can later be compared against data from diabetic patients to spot early signs of ulcer risk. Isolation Forest detects subtle distributed anomalies while KNN/LOF highlights sharper extremes, and the two methods show only partial overlap under a shared 5 percent contamination assumption. A modest correlation between the pressure and temperature signals supports using both sensors together. The resulting analytical pipeline is offered as a methodological foundation rather than a ready diagnostic, pending direct clinical validation on patients.

Core claim

The authors claim that applying Isolation Forest and KNN/LOF to 93,790 multi-sensor readings from 312 healthy-subject sessions successfully defines normal foot temperature and pressure patterns under a 5 percent contamination model, with Isolation Forest proving more sensitive to subtle anomalies and the overall framework providing a ready baseline for future studies that can directly link detected deviations to diabetic foot ulcer pathophysiology.

What carries the argument

Isolation Forest and Local Outlier Factor algorithms applied to paired temperature and pressure time-series data from wearable foot sensors under a fixed 5 percent contamination assumption.

If this is right

  • Isolation Forest identifies subtle distributed anomalies more readily than KNN/LOF.
  • KNN/LOF flags concentrated extreme deviations but labels a higher share of sessions as anomalous.
  • A mild positive correlation of 0.41-0.48 between pressure and temperature features supports combined multi-modal monitoring.
  • Differences between the algorithms reflect relative sensitivity under the shared contamination assumption rather than confirmed error rates.
  • The pipeline is positioned as a methodological foundation ready for direct clinical validation in diabetic cohorts.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Longitudinal follow-up of patients wearing the same sensors could test whether baseline anomalies predict actual ulcer formation.
  • The unsupervised baseline could later be paired with supervised classifiers once patient labels become available.
  • Similar reference models built from healthy data might apply to other biomechanical or chronic-disease monitoring tasks.
  • Varying the contamination rate to match known clinical prevalence could improve calibration for real-world deployment.

Load-bearing premise

Statistical outliers identified in healthy-subject data under a 5 percent contamination assumption will correspond to clinically relevant diabetic foot ulcer pathophysiology once the same methods are applied to patient measurements.

What would settle it

Collecting equivalent sensor data from diabetic patients with and without foot ulcers and checking whether the rate or type of detected anomalies correlates with clinical DFU indicators or ulcer incidence would directly test the claimed mapping.

Figures

Figures reproduced from arXiv: 2603.12278 by Md Tanvir Hasan Turja.

Figure 1
Figure 1. Figure 1: Visualization of Selected Patient IDs Based on Pressure Data [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Distribution of features across pData(Pressure Data) types [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Feature Importance for Pressure Anomaly Detection for Isolation [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Pressure Data (Anomaly Scores for Pressure Data) [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Concentration of Anomalies for some most important features [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: KNN Anomaly Detection Results (Pressure Data) [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Concentration of Anomalies for some most important features [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Visualization of Selected Capture IDs Based on Temperature Data [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Distribution of features across tData(Temperature data) types [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Feature Importance for Temperature Anomaly Detection for Iso [PITH_FULL_IMAGE:figures/full_fig_p019_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Temperature Data (Anomaly Scores for Temperature Data) [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Concentration of Anomalies for some most important features [PITH_FULL_IMAGE:figures/full_fig_p021_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: KNN Anomaly Detection Results (Temperature Data) [PITH_FULL_IMAGE:figures/full_fig_p022_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Concentration of Anomalies for some most important features [PITH_FULL_IMAGE:figures/full_fig_p022_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Heatmaps of Pressure Metrics by Month Heatmaps of Pressure Metrics by Month Heatmap analysis of pres￾sure metrics from October 2023 to June 2024 showed significant monthly variations. Maximum pressure peaked in January and June 2024, while ex￾tended step periods indicated potential risk in November 2023 and February 23 [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Heatmaps of Temperature Metrics by Month [PITH_FULL_IMAGE:figures/full_fig_p024_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Comparison of Anomaly Detection Frequencies Across Sensors [PITH_FULL_IMAGE:figures/full_fig_p025_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Correlation matrix between Pressure and Temperature features [PITH_FULL_IMAGE:figures/full_fig_p025_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Seasonal temperature trend Seasonal temperature trend * The analysis of foot temperature pat￾terns showed a significant drop in December 2023, potentially indicating reduced circulation or vasoconstriction due to colder weather, which could increase the risk of foot ulcers for diabetic patients. 26 [PITH_FULL_IMAGE:figures/full_fig_p026_19.png] view at source ↗
read the original abstract

Diabetic foot ulcers (DFUs) are a severe complication of diabetes associated with significant morbidity, amputation risk, and healthcare burden. Developing effective continuous monitoring frameworks requires first establishing reliable baseline models of normal foot biomechanics. This paper presents a feasibility study of an anomaly detection framework applied to time-series data from wearable foot sensors, specifically NTC thin-film thermocouples for temperature and FlexiForce A401 pressure sensors for plantar load monitoring. Data were collected from healthy adult subjects across 312 capture sessions on an instrumented pathway, generating 93,790 valid multi-sensor readings spanning September 2023 to June 2024. Two unsupervised algorithms, Isolation Forest and K-Nearest Neighbors using Local Outlier Factor (KNN/LOF), were applied to detect statistical deviations in foot temperature and pressure signals. Results show that Isolation Forest is more sensitive to subtle, distributed anomalies, while KNN/LOF identifies concentrated extreme deviations but flags a higher proportion of sessions not corroborated by Isolation Forest. Since no clinical ground truth is available, this difference is interpreted as lower specificity under the shared 5 percent contamination assumption rather than a confirmed false-positive rate. A mild positive correlation (0.41-0.48) between pressure and temperature features supports the case for combined multi-modal monitoring. These findings establish a validated baseline analytical pipeline and provide a methodological foundation for future clinical validation studies involving diabetic patients, where the relationship between detected anomalies and DFU-related pathophysiology can be directly assessed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 3 minor

Summary. The paper presents a feasibility study applying unsupervised anomaly detection (Isolation Forest and KNN/LOF) to multi-modal wearable foot sensor data (NTC temperature and FlexiForce pressure) collected from 312 sessions of healthy adult subjects, yielding 93,790 readings. It reports algorithm-specific anomaly patterns under a fixed 5% contamination assumption, a mild positive correlation (0.41-0.48) between pressure and temperature features, and positions the work as an empirical baseline pipeline to support future clinical validation in diabetic patients for DFU prevention.

Significance. If the pipeline holds under clinical extension, the work supplies a large, well-described healthy-subject dataset and reproducible unsupervised methods that can serve as a methodological foundation for multi-modal DFU monitoring frameworks. The explicit scoping to baseline establishment (rather than clinical mapping) and acknowledgment of absent ground truth are strengths that make the contribution proportionate and useful for subsequent studies.

major comments (1)
  1. [Abstract] Abstract and Results: the post-hoc attribution of KNN/LOF's higher session-level disagreement to 'lower specificity' under the shared 5% contamination assumption is not quantitatively tested; a sensitivity sweep over contamination rates (or comparison against a supervised proxy) would be needed to support this interpretation as more than descriptive.
minor comments (3)
  1. [Methods] Methods: explicit values for Isolation Forest hyperparameters (n_estimators, max_samples) and LOF k-neighbor count should be stated so the pipeline is fully reproducible from the 93,790-reading dataset.
  2. [Results] Results: the reported correlation range (0.41-0.48) should be accompanied by a p-value or confidence interval and by the exact feature definitions (e.g., mean vs. peak pressure) used in the calculation.
  3. [Discussion] Discussion: the claim that the pipeline provides a 'validated baseline' could be rephrased to 'empirically characterized baseline' to avoid implying external validation that the healthy-subject design cannot supply.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive feedback and positive assessment of our feasibility study as a methodological baseline. We address the single major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract and Results: the post-hoc attribution of KNN/LOF's higher session-level disagreement to 'lower specificity' under the shared 5% contamination assumption is not quantitatively tested; a sensitivity sweep over contamination rates (or comparison against a supervised proxy) would be needed to support this interpretation as more than descriptive.

    Authors: We agree that the current interpretation of algorithm-specific disagreement would benefit from quantitative support. Because the study is unsupervised and lacks ground-truth labels, a supervised proxy is not feasible; however, we will add a sensitivity sweep over contamination rates (1%, 3%, 5%, 7%, 10%) for both Isolation Forest and KNN/LOF. The revised Results section will report how session-level anomaly flags and inter-algorithm disagreement change across these rates, allowing readers to evaluate the robustness of the specificity claim under varying assumptions. Corresponding updates will be made to the Abstract. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper is a purely empirical feasibility study that applies standard off-the-shelf unsupervised algorithms (Isolation Forest, KNN/LOF) to raw sensor time-series collected from healthy subjects. No equations, predictions, or derived quantities are defined in terms of themselves; the 5% contamination assumption is stated explicitly as an input rather than fitted from the target outputs, and the reported correlation (0.41-0.48) is a direct statistical summary of the processed features. Conclusions are scoped to establishing a baseline pipeline with explicit acknowledgment of absent ground truth, so no load-bearing step reduces to a self-citation, ansatz, or renaming of the input data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the standard unsupervised anomaly detection premise that a fixed small fraction of observations are outliers and that these outliers will later map to clinical risk.

free parameters (1)
  • contamination rate = 0.05
    Fixed at 5% for both Isolation Forest and KNN/LOF to set the anomaly threshold.
axioms (1)
  • domain assumption Unsupervised statistical deviation detection on healthy data will identify patterns relevant to future diabetic pathology
    Invoked when interpreting algorithm outputs as a baseline for DFU studies without labeled validation data.

pith-pipeline@v0.9.0 · 5574 in / 1213 out tokens · 28147 ms · 2026-05-15T18:44:03.884299+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages

  1. [1]

    Carmine Gazzaruso, Pietro Gallotti, Arturo Pujia, Tiziana Montalcini, Andrea Giustina, and Adriana Coppola. Predictors of healing, ulcer re- currence and persistence, amputation and mortality in type 2 diabetic patients with diabetic foot: a 10-year retrospective cohort study.En- docrine, 71(1):59–68, 2021

    work page 2021

  2. [2]

    Caitlin W Hicks, Joseph K Canner, Nestoras Mathioudakis, Christopher Lippincott, Ronald L Sherman, and Christopher J Abularrage. Incidence and risk factors associated with ulcer recurrence among patients with diabetic foot ulcers treated in a multidisciplinary setting.Journal of Surgical Research, 246:243–250, 2020

    work page 2020

  3. [3]

    Reap- praising diabetic foot ulcers: a focus on mechanisms of ulceration and 31 clinical evaluation.The international journal of lower extremity wounds, 21(3):294–302, 2022

    Paola C Aldana, Alexander M Cartron, and Amor Khachemoune. Reap- praising diabetic foot ulcers: a focus on mechanisms of ulceration and 31 clinical evaluation.The international journal of lower extremity wounds, 21(3):294–302, 2022

    work page 2022

  4. [4]

    A review of wearable sensor systems to monitor plantar loading in the assessment of diabetic foot ulcers.IEEE Transactions on Biomedical Engineering, 67(7):1989– 2004, 2019

    Lefan Wang, Dominic Jones, Graham J Chapman, Heidi J Siddle, David A Russell, Ali Alazmani, and Peter Culmer. A review of wearable sensor systems to monitor plantar loading in the assessment of diabetic foot ulcers.IEEE Transactions on Biomedical Engineering, 67(7):1989– 2004, 2019

    work page 1989

  5. [5]

    Diagnosis and management of diabetic foot complications.Diabetes Care, 41(Supplement 1):S145–S150, 2018

    Andrew JM Boulton, David G Armstrong, Robert S Kirsner, Christo- pher E Attinger, Lawrence A Lavery, Benjamin A Lipsky, Joseph L Mills Sr, and John S Steinberg. Diagnosis and management of diabetic foot complications.Diabetes Care, 41(Supplement 1):S145–S150, 2018

    work page 2018

  6. [6]

    Stepping forward: a scoping systematic literature review on the health outcomes of smart sensor technologies for diabetic foot ulcers.Sensors, 24(6):2009, 2024

    Ioulietta Lazarou, Vasiliki Fiska, Lampros Mpaltadoros, Dimitris Tsaopoulos, Thanos G Stavropoulos, Spiros Nikolopoulos, George E Dafoulas, Zoe Dailiana, Alexandra Bargiota, and Ioannis Kompatsiaris. Stepping forward: a scoping systematic literature review on the health outcomes of smart sensor technologies for diabetic foot ulcers.Sensors, 24(6):2009, 2024

    work page 2009

  7. [7]

    The cost of diabetic foot ulcers and amputations to the national health service in england.Diabetic Medicine, 36(8):995–1002, 2019

    M Kerr, E Barron, P Chadwick, Timothy Evans, WM Kong, G Rayman, M Sutton-Smith, G Todd, B Young, and WJ Jeffcoate. The cost of diabetic foot ulcers and amputations to the national health service in england.Diabetic Medicine, 36(8):995–1002, 2019. 32

    work page 2019

  8. [8]

    Wound assessment, imaging and monitoring systems in diabetic foot ulcers: A systematic review

    Kai Siang Chan and Zhiwen Joseph Lo. Wound assessment, imaging and monitoring systems in diabetic foot ulcers: A systematic review. International wound journal, 17(6):1909–1923, 2020

    work page 1909

  9. [9]

    Assessment methods for patient- reported outcomes.Disease Management & Health Outcomes, 11(2):77– 86, 2003

    Michael R Hufford and Saul Shiffman. Assessment methods for patient- reported outcomes.Disease Management & Health Outcomes, 11(2):77– 86, 2003

    work page 2003

  10. [10]

    Zhongqiang Zhu, Lu Wang, Yaou Peng, Xiaoying Chu, Liyang Zhou, Yingying Jin, Hanwen Guo, Qiang Gao, Jianhua Yang, Xinyi Wang, et al. Continuous self-oxygenated double-layered hydrogel under natural light for real-time infection monitoring, enhanced photodynamic ther- apy, and hypoxia relief in refractory diabetic wounds healing.Advanced Functional Materia...

    work page 2022

  11. [11]

    Progress and challenges in fabrication of wearable sensors for health monitoring.Sensors and Actuators A: Physical, 312:112105, 2020

    Sara Nasiri and Mohammad Reza Khosravani. Progress and challenges in fabrication of wearable sensors for health monitoring.Sensors and Actuators A: Physical, 312:112105, 2020

    work page 2020

  12. [12]

    Review on wearables to monitor foot temperature in diabetic patients.Sensors, 19(4):776, 2019

    Jes´ us Mart´ ın-Vaquero, Ascensi´ on Hern´ andez Encinas, Araceli Queiruga- Dios, Juan Jos´ e Bull´ on, Alfonso Mart´ ınez-Nova, Jose Torre- blanca Gonz´ alez, and Cristina Bull´ on-Carbajo. Review on wearables to monitor foot temperature in diabetic patients.Sensors, 19(4):776, 2019. 33

    work page 2019

  13. [13]

    Towards the in- ternet of smart clothing: A review on iot wearables and garments for creating intelligent connected e-textiles.Electronics, 7(12):405, 2018

    Tiago M Fern´ andez-Caram´ es and Paula Fraga-Lamas. Towards the in- ternet of smart clothing: A review on iot wearables and garments for creating intelligent connected e-textiles.Electronics, 7(12):405, 2018

    work page 2018

  14. [14]

    The role of machine learning in advancing diabetic foot: a review.Frontiers in Endocrinology, 15:1325434, 2024

    Huifang Guan, Ying Wang, Ping Niu, Yuxin Zhang, Yanjiao Zhang, Runyu Miao, Xinyi Fang, Ruiyang Yin, Shuang Zhao, Jun Liu, et al. The role of machine learning in advancing diabetic foot: a review.Frontiers in Endocrinology, 15:1325434, 2024

    work page 2024

  15. [15]

    Unilateral remote temperature monitoring to predict future ulceration for the diabetic foot in remission.BMJ Open Diabetes Research & Care, 7(1), 2019

    Lawrence A Lavery, Brian J Petersen, David R Linders, Jonathan D Bloom, Gary M Rothenberg, and David G Armstrong. Unilateral remote temperature monitoring to predict future ulceration for the diabetic foot in remission.BMJ Open Diabetes Research & Care, 7(1), 2019

    work page 2019

  16. [16]

    Sicco A Bus, B Wouter, Jeff G van Baal, Tessa E Busch-Westbroek, Frans Nollet, and Jaap J van Netten. Effectiveness of at-home skin temperature monitoring in reducing the incidence of foot ulcer recur- rence in people with diabetes: a multicenter randomized controlled trial (diatemp).BMJ open diabetes research & care, 9(1), 2021

    work page 2021

  17. [17]

    Biomechanics of the diabetic foot: the road to foot ulceration

    Carine HM van Schie and Andrew JM Boulton. Biomechanics of the diabetic foot: the road to foot ulceration. InThe diabetic foot, pages 185–200. Springer, 2006

    work page 2006

  18. [18]

    Physiological state recognition via hrv 34 and fractal analysis using ai and unsupervised clustering.Information, 16(9):718, 2025

    Galya Georgieva-Tsaneva, Krasimir Cheshmedzhiev, Yoan-Aleksandar Tsanev, and Miroslav Dechev. Physiological state recognition via hrv 34 and fractal analysis using ai and unsupervised clustering.Information, 16(9):718, 2025

    work page 2025

  19. [19]

    An embedded wearable device for monitoring diabetic foot ulcer parameters

    Vishwajit V Kulkarni, Fatjon Seraj, and Jansi Rani Rathikumar. An embedded wearable device for monitoring diabetic foot ulcer parameters. InProceedings of the 13th ACM International Conference on PErvasive Technologies Related to Assistive Environments, pages 1–7, 2020

    work page 2020

  20. [20]

    Harnessing digital health technolo- gies to remotely manage diabetic foot syndrome: a narrative review

    Bijan Najafi and Ramkinker Mishra. Harnessing digital health technolo- gies to remotely manage diabetic foot syndrome: a narrative review. Medicina, 57(4):377, 2021

    work page 2021

  21. [21]

    Deep learning technologies for time series anomaly detection in healthcare: A review.Ieee Access, 11:117788–117799, 2023

    Xue Yang, Xuejun Qi, and Xiaobo Zhou. Deep learning technologies for time series anomaly detection in healthcare: A review.Ieee Access, 11:117788–117799, 2023

    work page 2023

  22. [22]

    Integrating technologies to enhance risk assessment for the early detection and prevention of pres- sure ulcers.Journal of Wound Care, 33(9):644–651, 2024

    Hannah Wilson, Pinar Avsar, Natalie McEvoy, Sorcha Byrne, Giulio Brunetti, Declan Patton, and Zena Moore. Integrating technologies to enhance risk assessment for the early detection and prevention of pres- sure ulcers.Journal of Wound Care, 33(9):644–651, 2024

    work page 2024

  23. [23]

    Impact of dataset size on classification performance: an empirical eval- uation in the medical domain.Applied sciences, 11(2):796, 2021

    Alhanoof Althnian, Duaa AlSaeed, Heyam Al-Baity, Amani Samha, Alanoud Bin Dris, Najla Alzakari, Afnan Abou Elwafa, and Heba Kurdi. Impact of dataset size on classification performance: an empirical eval- uation in the medical domain.Applied sciences, 11(2):796, 2021. 35

    work page 2021

  24. [24]

    Mobile and wearable sensors for data-driven health monitoring system: State-of-the-art and future prospect.Expert Systems with Applications, 202:117362, 2022

    Chioma Virginia Anikwe, Henry Friday Nweke, Anayo Chukwu Ikegwu, Chukwunonso Adolphus Egwuonwu, Fergus Uchenna Onu, Uzoma Rita Alo, and Ying Wah Teh. Mobile and wearable sensors for data-driven health monitoring system: State-of-the-art and future prospect.Expert Systems with Applications, 202:117362, 2022

    work page 2022

  25. [25]

    A comprehensive review of methods based on deep learning for diabetes-related foot ulcers.Frontiers in Endocrinology, 13:945020, 2022

    Jianglin Zhang, Yue Qiu, Li Peng, Qiuhong Zhou, Zheng Wang, and Min Qi. A comprehensive review of methods based on deep learning for diabetes-related foot ulcers.Frontiers in Endocrinology, 13:945020, 2022

    work page 2022

  26. [26]

    Chronic ulcers healing prediction through machine learning approaches: Preliminary results on diabetic foot ulcers case study.Journal of Clinical Medicine, 14(9):2943, 2025

    Elisabetta Spinazzola, Guillaume Picaud, Sara Becchi, Monica Pittarello, Elia Ricci, Marc Chaumont, G´ erard Subsol, Fabio Pareschi, Luc Teot, and Jacopo Secco. Chronic ulcers healing prediction through machine learning approaches: Preliminary results on diabetic foot ulcers case study.Journal of Clinical Medicine, 14(9):2943, 2025

    work page 2025

  27. [27]

    Context-awareness for mobile sensing: A survey and future directions.IEEE Communications Surveys & Tu- torials, 18(1):68–93, 2014

    ¨Ozg¨ ur Y¨ ur¨ ur, Chi Harold Liu, Zhengguo Sheng, Victor CM Leung, Wil- frido Moreno, and Kin K Leung. Context-awareness for mobile sensing: A survey and future directions.IEEE Communications Surveys & Tu- torials, 18(1):68–93, 2014

    work page 2014

  28. [28]

    Iot based wearable smart insole

    Jochen Wilden, Asti Chandrakar, Atul Ashok, and Neeli Prasad. Iot based wearable smart insole. In2017 Global Wireless Summit (GWS), pages 186–192. IEEE, 2017. 36

    work page 2017