Edge-Enabled Pattern Recognition Architecture for Energy-Efficient Intelligent Computing in Next-Generation Wearable Health Devices
DOI:
https://doi.org/10.59461/ijdiic.v5i2.283Keywords:
Spiking neural networks, Wearable health devices, Edge computing, Anomaly detection, Energy-efficient inferenceAbstract
Wearable health devices have moved from step-counting toys to genuine medical instruments, but the gap between what they can sense and what they can decide on-device is still wide. Most clinical-grade pattern recognition still happens in the cloud, which costs latency, privacy, and battery life. In this paper, we propose NeuroPulse-Edge, a lightweight on-device architecture that combines spiking neurons with a spike-only attention block to deliver real-time anomaly detection on wearable biosignals at a sub-2 mW power budget. The system encodes ECG, PPG, accelerometer, and skin-temperature streams into spike trains through a delta-modulation front end, runs them through a stack of leaky integrate-and-fire neurons, a spike convolution block, and a binary spiking-attention block, and ends in a small INT8 classifier. The attention block is the key piece — it preserves the long-range temporal sensitivity that arrhythmia recognition needs, but does it with logical AND-OR and popcount instead of floating-point multiply-accumulates, so it costs almost nothing to run on a microcontroller. We evaluate the system on four 2025–2026 wearable benchmarks (CACHET-CADB, WildPPG, the 2025 multimodal stress dataset, and a 2026 long-term smartwatch arrhythmia release) against five strong baselines. NeuroPulse-Edge reaches 92.1% accuracy and a 0.921 macro F1 on a five-class arrhythmia task, draws 1.17 mW continuously on an ARM Cortex-M4 testbench, and answers in 14 ms per inference — an 8.2× power reduction and a 4.4× latency reduction over the strongest deep-learning baseline at the same accuracy, with every gain confirmed at p < 0.01 (Wilcoxon signed-rank with Holm–Bonferroni correction). Embedded analyses establish a surrogate-gradient convergence bound and a spike-sparsity energy bound for the architecture. The findings indicate that spike-based attention is a practical building block for clinical-grade wearable AI.
References
J. Sink et al., “Correlation between high- and low-voltage impedance measurements following subcutaneous implantable cardioverter-defibrillator implantation,” Hear Rhythm, vol. 21, no. 4, pp. 492–494, Apr. 2024, doi: 10.1016/j.hrthm.2023.12.018.
M. Nalubega Moses H.F., “Wearable Health Devices and Data Analytics: Trends and Insights,” Res Invent J Biol Appl Sci, vol. 5, no. 2, pp. 29–32, Feb. 2025, doi: 10.59298/RIJBAS/2025/522932.
D. P. Judge et al., “Efficacy of Acoramidis on All-Cause Mortality and Cardiovascular Hospitalization in Transthyretin Amyloid Cardiomyopathy,” J Am Coll Cardiol, vol. 85, no. 10, pp. 1003–1014, Mar. 2025, doi: 10.1016/j.jacc.2024.11.042.
J. F. Sanchez, S. P. Gaine, and K. N. Aronis, “Antiarrhythmic Medications for Acute Management of Ventricular Arrhythmias,” Card Electrophysiol Clin, vol. 18, no. 1, pp. 79–89, Mar. 2026, doi: 10.1016/j.ccep.2025.10.005.
B. Pokharel, D. B. Thiyam, S. Devkota, and D. K. Sah, “Development and Prototyping of Oxygen Analyzer,” in ECSA-11, Basel Switzerland: MDPI, Nov. 2024, p. 82. doi: 10.3390/ecsa-11-20447.
R. Zhang et al., “HPSpeech: Silent Speech Interface for Commodity Headphones,” in Proceedings of the 2023 International Symposium on Wearable Computers, New York, NY, USA: ACM, Oct. 2023, pp. 60–65. doi: 10.1145/3594738.3611365.
M. Alghieth, “DeepECG-Net: a hybrid transformer-based deep learning model for real-time ECG anomaly detection,” Sci Rep, vol. 15, no. 1, p. 20714, Jul. 2025, doi: 10.1038/s41598-025-07781-1.
P. Busia, M. A. Scrugli, V. J.-B. Jung, L. Benini, and P. Meloni, “A Tiny Transformer for Low-Power Arrhythmia Classification on Microcontrollers,” IEEE Trans Biomed Circuits Syst, vol. 19, no. 1, pp. 142–152, Feb. 2025, doi: 10.1109/TBCAS.2024.3401858.
T. Huang et al., “DMC-LIBSAS: A Laser-Induced Breakdown Spectroscopy Analysis System with Double-Multi Convolutional Neural Network for Accurate Traceability of Chinese Medicinal Materials,” Sensors, vol. 25, no. 7, p. 2104, Mar. 2025, doi: 10.3390/s25072104.
P. P. Ray, “A review on TinyML: State-of-the-art and prospects,” J King Saud Univ - Comput Inf Sci, vol. 34, no. 4, pp. 1595–1623, Apr. 2022, doi: 10.1016/j.jksuci.2021.11.019.
M. Tri Lê, P. Wolinski, and J. Arbel, “Efficient Neural Networks for Tiny Machine Learning: A Comprehensive Review,” ACM Trans Intell Syst Technol, vol. 17, no. 4, pp. 1–41, Aug. 2026, doi: 10.1145/3798276.
M. Hizem, M. O.-E. Aoueileyine, S. B. Belhaouari, A. EL Omri, and R. Bouallegue, “Sustainable E-Health: Energy-Efficient Tiny AI for Epileptic Seizure Detection via EEG,” Biomed Eng Comput Biol, vol. 16, Aug. 2025, doi: 10.1177/11795972241283101.
L. Yuan, J. Wei, and Y. Liu, “Spiking neural networks for EEG signal analysis using wavelet transform,” Front Neurosci, vol. 19, Oct. 2025, doi: 10.3389/fnins.2025.1652274.
Y. Li, R. Yin, Y. Kim, and P. Panda, “Efficient human activity recognition with spatio-temporal spiking neural networks,” Front Neurosci, vol. 17, Sep. 2023, doi: 10.3389/fnins.2023.1233037.
E. Kim and Y. Kim, “Exploring the potential of spiking neural networks in biomedical applications: advantages, limitations, and future perspectives,” Biomed Eng Lett, vol. 14, no. 5, pp. 967–980, Sep. 2024, doi: 10.1007/s13534-024-00403-1.
S. Alinsaif, “Neuromorphic Computing for Long-Term Cardiac Health: A Review of Spiking Neural Networks in Low-Power Wearable Electronics,” Electronics, vol. 15, no. 6, p. 1179, Mar. 2026, doi: 10.3390/electronics15061179.
S. Davidson and S. B. Furber, “Comparison of Artificial and Spiking Neural Networks on Digital Hardware,” Front Neurosci, vol. 15, Apr. 2021, doi: 10.3389/fnins.2021.651141.
L.-C. Duan, N.-N. Minh, T.-C. Dung, and T.-T.-T. Linh, “Energy-Efficient TinyML Approach for Wearable Fall Detection on Edge Devices Using Spatial-Temporal Deep Learning,” Int J Technol, vol. 17, no. 3, p. 919, May 2026, doi: 10.14716/ijtech.v17i3.8445.
Z. Zhou et al., “Spikformer: When Spiking Neural Network Meets Transformer,” Nov. 2022, doi: 10.48550/arXiv.2209.15425.
Y. She, “A Robust PPO-optimized Tabular Transformer Framework for Intrusion Detection in Industrial IoT Systems,” May 2025.
B. Ayasi, C. J. Carmona, M. Saleh, and A. M. García-Vico, “A Practical Tutorial on Spiking Neural Networks: Comprehensive Review, Models, Experiments, Software Tools, and Implementation Guidelines,” Eng, vol. 6, no. 11, p. 304, Nov. 2025, doi: 10.3390/eng6110304.
G. Indiveri and S.-C. Liu, “Memory and Information Processing in Neuromorphic Systems,” Proc IEEE, vol. 103, no. 8, pp. 1379–1397, Aug. 2015, doi: 10.1109/JPROC.2015.2444094.
M. Davies et al., “Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook,” Proc IEEE, vol. 109, no. 5, pp. 911–934, May 2021, doi: 10.1109/JPROC.2021.3067593.
N. L. Kazanskiy, P. A. Khorin, and S. N. Khonina, “Biochips on the Move: Emerging Trends in Wearable and Implantable Lab-on-Chip Health Monitors,” Electronics, vol. 14, no. 16, p. 3224, Aug. 2025, doi: 10.3390/electronics14163224.
S. G. Dhekane, H. Haresamudram, M. Thukral, and T. Plötz, “How Much Unlabeled Data is Really Needed for Effective Self-Supervised Human Activity Recognition?,” in Proceedings of the 2023 International Symposium on Wearable Computers, New York, NY, USA: ACM, Oct. 2023, pp. 66–70. doi: 10.1145/3594738.3611366.
E. O. Neftci, H. Mostafa, and F. Zenke, “Surrogate Gradient Learning in Spiking Neural Networks: Bringing the Power of Gradient-Based Optimization to Spiking Neural Networks,” IEEE Signal Process Mag, vol. 36, no. 6, pp. 51–63, Nov. 2019, doi: 10.1109/MSP.2019.2931595.
W. Fang et al., “SpikingJelly: An open-source machine learning infrastructure platform for spike-based intelligence,” Sci Adv, vol. 9, no. 40, Oct. 2023, doi: 10.1126/sciadv.adi1480.
A. Mehrabi, N. Sreenivasan, U. Gunawardana, and G. Gargiulo, “Hybrid Spike-Encoded Spiking Neural Networks for Real-Time EEG Seizure Detection: A Comparative Benchmark,” Biomimetics, vol. 11, no. 1, p. 75, Jan. 2026, doi: 10.3390/biomimetics11010075.
S. Ikram et al., “Transformer-based ECG classification for early detection of cardiac arrhythmias,” Front Med, vol. 12, Aug. 2025, doi: 10.3389/fmed.2025.1600855.
R. Wang et al., “TAC-ECG: A task-adaptive classification method for electrocardiogram based on cross-modal contrastive learning and low-rank convolutional adapter,” Comput Methods Programs Biomed, vol. 270, p. 108918, Oct. 2025, doi: 10.1016/j.cmpb.2025.108918.
J. B. Moody et al., “A Foundation Transformer Model with Self-Supervised Learning for ECG-Based Assessment of Cardiac and Coronary Function,” NEJM AI, vol. 2, no. 12, Nov. 2025, doi: 10.1056/AIoa2500164.
H. Wu, C. Chen, and K. Weng, “An Energy-Efficient Strategy for Microcontrollers,” Appl Sci, vol. 11, no. 6, p. 2581, Mar. 2021, doi: 10.3390/app11062581.
D. T. Pham, M. K. Titkanlou, and R. Mouček, “A hybrid Spiking Neural Network–Transformer architecture for motor imagery and sleep apnea detection,” Front Neurosci, vol. 19, Dec. 2025, doi: 10.3389/fnins.2025.1716204.
L. Almeida, R. Teixeira, G. Baldoni, M. Antunes, and R. L. Aguiar, “Federated Learning for a Dynamic Edge: A Modular and Resilient Approach,” Sensors, vol. 25, no. 12, p. 3812, Jun. 2025, doi: 10.3390/s25123812.
H. Sherief, M. Fayik, A. Abd El-Latief, M. Naim Anwar, A. M. Tawfik, and A. Elsayed, “The effect of the surface recombination velocity on 2D thermoelastic semiconductor solid sphere problem,” Alexandria Eng J, vol. 127, pp. 1126–1142, Aug. 2025, doi: 10.1016/j.aej.2025.07.012.
A. R. Keivanimehr and M. Akbari, “TinyML and edge intelligence applications in cardiovascular disease: A survey,” Comput Biol Med, vol. 186, p. 109653, Mar. 2025, doi: 10.1016/j.compbiomed.2025.109653.
D. Kumar, S. Puthusserypady, H. Dominguez, K. Sharma, and J. E. Bardram, “CACHET-CADB: A Contextualized Ambulatory Electrocardiography Arrhythmia Dataset,” Front Cardiovasc Med, vol. 9, Jul. 2022, doi: 10.3389/fcvm.2022.893090.
M. Meier, B. U. Demirel, and C. Holz, “WildPPG: A Real-World PPG Dataset of Long Continuous Recordings,” Dec. 2024, doi: 10.3389/fcvm.2022.893981.
S. Park, D. Zheng, and U. Lee, “A PPG Signal Dataset Collected in Semi-Naturalistic Settings Using Galaxy Watch,” Sci Data, vol. 12, no. 1, p. 892, May 2025, doi: 10.1038/s41597-025-05152-z.
D. Han et al., "Multiclass arrhythmia classification using multimodal smartwatch photoplethysmography signals collected in real-life settings," December 13, 2024. doi: 10.21203/rs.3.rs-5463126/v1.
M. Hosseini et al., “A multimodal stress detection dataset with facial expressions and physiological signals,” Sci Data, vol. 12, no. 1, p. 1844, Nov. 2025, doi: 10.1038/s41597-025-05812-0.
A. N. Burkitt, “A review of the integrate-and-fire neuron model: I. Homogeneous synaptic input.,” Biol Cybern, vol. 95, no. 1, pp. 1–19, Jul. 2006, doi: 10.1007/s00422-006-0068-6.
Published
Issue
Section
License
Copyright (c) 2026 Anirudha Gaikwad, Atit Gaikwad, Dheeraj Lokhande
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
