Aprendizaje automático integrado de bajo consumo en microcontroladores: revisión bibliográfica sistemática
DOI:
https://doi.org/10.15649/2346030X.4183Palabras clave:
machine learning, low power, microcontrollers, tinyml, Literature ReviewResumen
La creciente adopción de la inteligencia artificial (IA) y el aprendizaje automático (ML) en los sistemas de ingeniería se ha basado tradicionalmente en el procesamiento basado en la nube y en plataformas de alto rendimiento, lo que genera importantes costes computacionales, energéticos y de infraestructura. La aparición del aprendizaje automático integrado de bajo consumo, conocido como TinyML, ha permitido el despliegue de modelos de ML directamente en microcontroladores (MCU) con recursos limitados, lo que ha permitido alcanzar una inteligencia cercana a la de los sensores, reducir la latencia y mejorar la eficiencia energética. Este artículo presenta una revisión bibliográfica sistemática sobre el uso de microcontroladores de bajo consumo para implementar técnicas de aprendizaje automático. El objetivo es sintetizar y analizar las investigaciones existentes sobre el aprendizaje automático integrado en MCU, haciendo hincapié en las plataformas de hardware, los ámbitos de aplicación y los modelos, métodos y algoritmos adoptados bajo estrictas restricciones de recursos. La revisión sigue la metodología estándar de revisión sistemática, utilizando la base de datos Scopus para publicaciones de 2020 a 2025. Los estudios analizados revelan que los microcontroladores basados en ARM Cortex-M y los dispositivos de clase Arduino son las plataformas predominantes. Las aplicaciones abarcan la monitorización medioambiental, los sistemas energéticos, la agricultura, la sanidad, la seguridad industrial, la biosensórica y las infraestructuras inteligentes. Los modelos más utilizados incluyen redes neuronales convolucionales ligeras, redes densas compactas y algoritmos clásicos de aprendizaje automático, optimizados mediante cuantificación, poda y compresión. Esta revisión destaca la creciente madurez del aprendizaje automático en microcontroladores de baja potencia e identifica las tendencias clave, las limitaciones y las compensaciones de diseño que dan forma a las implementaciones actuales de ML integrado.
Referencias
[1] R. Kallimani, K. Pai, P. Raghuwanshi, S. Iyer, and O. L. A. López, “TinyML: Tools, Applications, Challenges, and Future Research Directions,” Multimed. Tools Appl., Mar. 2023, doi: 10.1007/s11042-023-16740-9.
[2] R. I. Mukhamediev et al., “Review of Artificial Intelligence and Machine Learning Technologies: Classification, Restrictions, Opportunities and Challenges,” Mathematics 2022, Vol. 10, Page 2552, vol. 10, no. 15, p. 2552, Jul. 2022, doi: 10.3390/MATH10152552.
[3] M. Soori, B. Arezoo, and R. Dastres, “Artificial intelligence, machine learning and deep learning in advanced robotics, a review,” Jan. 01, 2023, KeAi Communications Co. doi: 10.1016/j.cogr.2023.04.001.
[4] M. Ali, A. Dewan, A. K. Sahu, and M. M. Taye, “Understanding of Machine Learning with Deep Learning: Architectures, Workflow, Applications and Future Directions,” Computers 2023, Vol. 12, Page 91, vol. 12, no. 5, p. 91, Apr. 2023, doi: 10.3390/COMPUTERS12050091.
[5] P. Zhou, “Study on CPU and FPGA in Artificial Intelligence and Machine Learning,” Proceedings - 2024 13th International Conference of Information and Communication Technology, ICTech 2024, pp. 462–466, 2024, doi: 10.1109/ICTECH63197.2024.00090.
[6] T. Baji, “Evolution of the GPU Device widely used in AI and Massive Parallel Processing,” 2018 IEEE Electron Devices Technology and Manufacturing Conference, EDTM 2018 - Proceedings, pp. 7–9, Jul. 2018, doi: 10.1109/EDTM.2018.8421507.
[7] L. Alzubaidi et al., “Review of deep learning: concepts, CNN architectures, challenges, applications, future directions,” Journal of Big Data 2021 8:1, vol. 8, no. 1, pp. 1–74, Mar. 2021, doi: 10.1186/S40537-021-00444-8.
[8] D. Gyawali, “Comparative Analysis of CPU and GPU Profiling for Deep Learning Models,” Sep. 2023, [Online]. Available: http://arxiv.org/abs/2309.02521.
[9] A. S. Al-Ahmad and H. Kahtan, “Cloud Computing Review: Features and Issues,” 2018 International Conference on Smart Computing and Electronic Enterprise, ICSCEE 2018, Nov. 2018, doi: 10.1109/ICSCEE.2018.8538387.
[10] Y. Zhao, H. Zhang, L. An, and Q. Liu, “Improving the approaches of traffic demand forecasting in the big data era,” Cities, vol. 82, pp. 19–26, Dec. 2018, doi: 10.1016/j.cities.2018.04.015.
[11] T. Staal, “The impact of the Internet of Things on the demand of cloud resources,” 2022. Accessed: Jun. 01, 2024. [Online]. Available: https://purl.utwente.nl/essays/91318.
[12] M. Satyanarayanan, “The Emergence of Edge Computing,” Computer (Long Beach. Calif)., vol. 50, no. 1, pp. 30–39, Jan. 2017, doi: 10.1109/MC.2017.9.
[13] F. C. Andriulo, M. Fiore, M. Mongiello, E. Traversa, and V. Zizzo, “Edge Computing and Cloud Computing for Internet of Things: A Review,” Informatics 2024, Vol. 11, Page 71, vol. 11, no. 4, p. 71, Sep. 2024, doi: 10.3390/INFORMATICS11040071.
[14] J. Lin, L. Zhu, W.-M. Chen, W.-C. Wang, and S. Han, “Tiny Machine Learning: Progress and Futures,” IEEE Circuits and Systems Magazine, pp. 8–34, Mar. 2024, doi: 10.1109/MCAS.2023.3302182.
[15] C. R. Banbury et al., “Benchmarking TinyML Systems: Challenges and Direction,” Mar. 2020, [Online]. Available: http://arxiv.org/abs/2003.04821.
[16] J. Sen et al., “Machine Learning: Algorithms, Models, and Applications,” Jan. 2022, doi: 10.5772/intechopen.94615.
[17] L. Heim, A. Biri, Z. Qu, and L. Thiele, “Measuring what Really Matters: Optimizing Neural Networks for TinyML,” Apr. 2021, [Online]. Available: http://arxiv.org/abs/2104.10645.
[18] S. S. Saha, S. S. Sandha, and M. Srivastava, “Machine Learning for Microcontroller-Class Hardware: A Review,” IEEE Sens. J., vol. 22, no. 22, pp. 21362–21390, Dec. 2022, doi: 10.1109/JSEN.2022.3210773.
[19] R. A. Mouha and R. A. Mouha, “Internet of Things (IoT),” Journal of Data Analysis and Information Processing, vol. 9, no. 2, pp. 77–101, Mar. 2021, doi: 10.4236/JDAIP.2021.92006.
[20] C. Willson Joseph and G. J. Willsie Kathrine, “Intelligent System with the IoT: A survey on techniques of Artificial Intelligence over the field of Internet of Things,” 8th International Conference on Advanced Computing and Communication Systems, ICACCS 2022, pp. 347–351, 2022, doi: 10.1109/ICACCS54159.2022.9785282.
[21] J. F. Burnham, “Scopus database: A review,” Mar. 08, 2006. doi: 10.1186/1742-5581-3-1.
[22] C. Manterola, P. Astudillo, E. Arias, and N. Claros, “Revisiones sistemáticas de la literatura. Qué se debe saber acerca de ellas,” Cir. Esp., vol. 91, no. 3, pp. 149–155, Mar. 2013, doi: 10.1016/J.CIRESP.2011.07.009.
[23] D. Evans, “The systematic review report,” Collegian, vol. 11, no. 2, pp. 8–11, Jan. 2004, doi: 10.1016/S1322-7696(08)60448-5.
[24] M. J. Grant and A. Booth, “A typology of reviews: An analysis of 14 review types and associated methodologies,” Health Info. Libr. J., vol. 26, no. 2, pp. 91–108, Jun. 2009, doi: 10.1111/J.1471-1842.2009.00848.X.
[25] R. Sanchez-Iborra and A. F. Skarmeta, “TinyML-Enabled Frugal Smart Objects: Challenges and Opportunities,” IEEE Circuits and Systems Magazine, vol. 20, no. 3, pp. 4–18, Jul. 2020, doi: 10.1109/MCAS.2020.3005467.
[26] P. P. Ray, “A review on TinyML: State-of-the-art and prospects,” Apr. 01, 2022, King Saud bin Abdulaziz University. doi: 10.1016/j.jksuci.2021.11.019.
[27] N. N. Alajlan and D. M. Ibrahim, “TinyML: Enabling of Inference Deep Learning Models on Ultra-Low-Power IoT Edge Devices for AI Applications,” Micromachines (Basel)., vol. 13, no. 6, Jun. 2022, doi: 10.3390/mi13060851.
[28] H. Han and J. Siebert, “TinyML: A Systematic Review and Synthesis of Existing Research,” 4th International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2022 - Proceedings, pp. 269–274, 2022, doi: 10.1109/ICAIIC54071.2022.9722636.
[29] B. Moons, D. Bankman, L. Yang, B. Murmann, M. Verhelst, and K. Leuven, “BinarEye: An Always-On Energy-Accuracy-Scalable Binary CNN Processor With All Memory On Chip In 28nm CMOS,” in IEEE Custom Integrated Circuits Conference (CICC), 2018. doi: 10.1109/CICC.2018.8357071.
[30] G. Signoretti, M. Silva, P. Andrade, I. Silva, E. Sisinni, and P. Ferrari, “An evolving tinyml compression algorithm for iot environments based on data eccentricity,” Sensors, vol. 21, no. 12, Jun. 2021, doi: 10.3390/s21124153.
[31] Z. Ksira, A. Mellit, N. Blasuttigh, and A. Massi Pavan, “A Novel Embedded System for Real-Time Fault Diagnosis of Photovoltaic Modules,” IEEE J. Photovolt., vol. 14, no. 2, pp. 354–362, Mar. 2024, doi: 10.1109/JPHOTOV.2024.3359462.
[32] L. S. Martinez-Rau, J. O. Chelotti, L. L. Giovanini, V. Adin, B. Oelmann, and S. Bader, “On-Device Feeding Behavior Analysis of Grazing Cattle,” IEEE Trans. Instrum. Meas., vol. 73, pp. 1–13, 2024, doi: 10.1109/TIM.2024.3376013.
[33] J. I. de Oliveira Filho, M. C. Faleiros, D. C. Ferreira, V. Mani, and K. N. Salama, “Empowering Electrochemical Biosensors with AI: Overcoming Interference for Precise Dopamine Detection in Complex Samples,” Advanced Intelligent Systems, vol. 5, no. 10, p. 2300227, Oct. 2023, doi: 10.1002/AISY.202300227.
[34] I. Cappelli et al., “Enhanced Visible Light Localization Based on Machine Learning and Optimized Fingerprinting in Wireless Sensor Networks,” IEEE Trans. Instrum. Meas., vol. 72, 2023, doi: 10.1109/TIM.2023.3240220.
[35] G. Peruzzi, A. Pozzebon, and M. Van Der Meer, “Fight Fire with Fire: Detecting Forest Fires with Embedded Machine Learning Models Dealing with Audio and Images on Low Power IoT Devices,” Sensors 2023, Vol. 23, Page 783, vol. 23, no. 2, p. 783, Jan. 2023, doi: 10.3390/S23020783.
[36] A. Mongardi et al., “Hand Gestures Recognition for Human-Machine Interfaces: A Low-Power Bio-Inspired Armband,” IEEE Trans. Biomed. Circuits Syst., vol. 16, no. 6, pp. 1348–1365, Dec. 2022, doi: 10.1109/TBCAS.2022.3211424.
[37] A. Roy, H. Dutta, H. Griffith, and S. Biswas, “An On-Device Learning System for Estimating Liquid Consumption from Consumer-Grade Water Bottles and Its Evaluation,” Sensors 2022, Vol. 22, Page 2514, vol. 22, no. 7, p. 2514, Mar. 2022, doi: 10.3390/S22072514.
[38] A. Krayden et al., “CMOS-MEMS Gas Sensor Dubbed GMOS for SelectiveAnalysis of Gases with Tiny Edge Machine Learning,” Engineering Proceedings 2022, Vol. 27, Page 81, vol. 27, no. 1, p. 81, Nov. 2022, doi: 10.3390/ECSA-9-13316.
[39] F. Dellragnola, U. Pale, R. Marino, A. Arza, and D. Atienza, “MBioTracker: Multimodal Self-Aware Bio-Monitoring Wearable System for Online Workload Detection,” IEEE Trans. Biomed. Circuits Syst., vol. 15, no. 5, pp. 994–1007, Oct. 2021, doi: 10.1109/TBCAS.2021.3110317.
[40] F. Shabani, H. Philamore, and F. Matsuno, “An energy-autonomous chemical oxygen demand sensor using a microbial fuel cell and embedded machine learning,” IEEE Access, vol. 9, pp. 108689–108701, 2021, doi: 10.1109/ACCESS.2021.3101496.
[41] N. Mohan, D. Abdelrahman, N. F. Ali, and M. Atef, “An Integrated High-Gain Wide-Dynamic Range Photoplethysmography Sensor for Cardiac Health Monitoring,” IEEE Sens. J., vol. 24, no. 7, pp. 10375–10384, Apr. 2024, doi: 10.1109/JSEN.2024.3367898.
[42] M. A. O. Zishan, H. M. Shihab, S. S. Islam, M. A. Riya, G. M. Rahman, and J. Noor, “Dense neural network based arrhythmia classification on low-cost and low-compute micro-controller,” Expert Syst. Appl., vol. 239, p. 122560, Apr. 2024, doi: 10.1016/J.ESWA.2023.122560.
[43] E. Tabanelli, D. Brunelli, A. Acquaviva, and L. Benini, “Trimming Feature Extraction and Inference for MCU-Based Edge NILM: A Systematic Approach,” IEEE Trans. Industr. Inform., vol. 18, no. 2, pp. 943–952, Feb. 2022, doi: 10.1109/TII.2021.3078186.
[44] J. Xiao et al., “ULECGNet: An Ultra-Lightweight End-to-End ECG Classification Neural Network,” IEEE J. Biomed. Health Inform., vol. 26, no. 1, pp. 206–217, Jan. 2022, doi: 10.1109/JBHI.2021.3090421.
[45] X. Wang, L. Cavigelli, T. Schneider, and L. Benini, “Sub-100 μW Multispectral Riemannian Classification for EEG-Based Brain–Machine Interfaces,” IEEE Trans. Biomed. Circuits Syst., vol. 15, no. 6, pp. 1149–1160, Dec. 2021, doi: 10.1109/TBCAS.2021.3137290.
[46] M. Zanghieri et al., “An Extreme-Edge TCN-Based Low-Latency Collision-Avoidance Safety System for Industrial Machinery,” IEEE Access, vol. 12, pp. 16009–16021, 2024, doi: 10.1109/ACCESS.2024.3357510.
[47] K. Xu, H. Zhang, Y. Li, Y. Zhang, R. Lai, and Y. Liu, “An Ultra-Low Power TinyML System for Real-Time Visual Processing at Edge,” IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 70, no. 7, pp. 2640–2644, Jul. 2023, doi: 10.1109/TCSII.2023.3239044.
[48] A. Albanese, M. Nardello, G. Fiacco, and D. Brunelli, “Tiny Machine Learning for High Accuracy Product Quality Inspection,” IEEE Sens. J., vol. 23, no. 2, pp. 1575–1583, Jan. 2023, doi: 10.1109/JSEN.2022.3225227.
[49] Y. Zhang, V. Adin, S. Bader, and B. Oelmann, “Leveraging Acoustic Emission and Machine Learning for Concrete Materials Damage Classification on Embedded Devices,” IEEE Trans. Instrum. Meas., vol. 72, 2023, doi: 10.1109/TIM.2023.3307751.
[50] E. Manor and S. Greenberg, “Custom Hardware Inference Accelerator for TensorFlow Lite for Microcontrollers,” IEEE Access, vol. 10, pp. 73484–73493, 2022, doi: 10.1109/ACCESS.2022.3189776.
Descargas
Publicado
Cómo citar
Descargas
Número
Sección
Licencia
Derechos de autor 2025 AiBi Revista de Investigación, Administración e Ingeniería
Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.
La revista ofrece acceso abierto bajo una Licencia Creative Commons Attibution License
Esta obra está bajo una licencia Creative Commons Attribution (CC BY 4.0).
