Blucher Engineering Proceedings

Setembro 2025, vol.12, num.1 - XXXII Simpósio Internacional de Engenharia
Trabalho completo - Open Access Idioma principal | Segundo principal

Abordagem Conceitual para Implementação de Redes Neurais LSTM em C para Prognóstico e Gerenciamento de Saúde de Veículos

Conceptual Approach to Implementing LSTM Neural Networks in C for Vehicle Prognostics and Health Management

FERNANDES, Thiago Bastos; PONTE, Leonardo Felipe Nunes; WATERMANN, Matheus Gomes Camargo; BORSATO, Milton

Trabalho completo:

A aplicação de redes Long Short-Term Memory (LSTM) em Prognóstico e Saúde de Máquina (PHM) tem ganhado destaque devido à sua eficácia em capturar dependências temporais em dados de séries temporais. No entanto, a implementação de modelos LSTM em sistemas automotivos embarcados apresenta desafios que vão além das implementações tradicionais. Este artigo explora um framework conceitual para a execução de inferência LSTM na linguagem de programação C, abordando as dificuldades de adaptação dos cálculos de redes neurais a ambientes com recursos limitados. O estudo apresenta um modelo LSTM leve, inicialmente desenvolvido e treinado em Python com um conjunto de dados sintético que simula a degradação de sistemas para predição da Vida Útil Remanescente (RUL). O modelo treinado foi então reimplementado em C e implantado em um microcontrolador. Todos os pesos treinados (de entrada, recorrentes e viés) foram exportados manualmente do Python e adaptados para uso em C, possibilitando inferência em tempo real sem dependências externas. Os resultados mostram alta consistência entre as plataformas, validando a implementação embarcada.

The application of Long Short-Term Memory (LSTM) networks in Prognostics and Health Management (PHM) has gained significant traction due to their effectiveness in capturing temporal dependencies in time-series data. However, deploying LSTM models on embedded automotive systems presents challenges that extend beyond traditional implementations. This paper explores a conceptual framework for implementing LSTM inference in the C programming language, addressing the difficulties of adapting neural network computations to resource-constrained environments. The study presents a lightweight LSTM model that was first developed and trained in Python on a synthetic time-series dataset simulating system degradation for Remaining Useful Life (RUL) prediction. The trained model was then reimplemented in the C programming language and deployed on a microcontroller. All trained weights (input, recurrent, and bias) were manually exported from Python and adapted for use in C, enabling real-time inference without external dependencies. Results show high consistency between platforms, validating the embedded implementation.

Download (PDF)

Palavras-chave: - -

DOI: 10.5151/simea2025-PAP18

Referências bibliográficas
  • [1] Aqui está sua lista de referências bibliográficas separada corretamente, sem espaçamento entre as linhas:
  • [2]
  • [3]
  • [4]
  • [5] AL-SAEDI, Ibtesam R. K.; MOHAMMED, Farag Mahel; OBAYES, Saifaldeen S. CNC machine based on embedded wireless and Internet of Things for workshop development. In: INTERNATIONAL CONFERENCE ON CONTROL, AUTOMATION AND DIAGNOSIS – ICCAD, 2017, Paris: IEEE, 2017. p. 439–444.
  • [6] ALLAL, Zaid; NOURA, Hassan N.; CHAHINE, Khaled. Efficient health indicators for the prediction of the remaining useful life of proton exchange membrane fuel cells. Energy Conversion and Management: X, v. 21, p. 100503, 2024.
  • [7] BENTSEN, Lars Ødegaard; WARAKAGODA, Narada Dilp; STENBRO, Roy; ENGELSTAD, Paal. A unified graph formulation for spatio-temporal wind forecasting. Energies, v. 16, n. 20, art. 7179, 2023. DOI: 10.3390/en16207179.
  • [8] CABRERA, Diego; VILLACÍS, Mauricio; CERRADA, Mariela; SÁNCHEZ, René-Vinicio; LI, Chuan; SANCHO, Fernando; LONG, Jianyu; ESTUPINAN, Edgar. Adversarial fault detector guided by one-class learning for a multistage centrifugal pump. IEEE/ASME Transactions on Mechatronics, v. 28, n. 3, p. 1395–1403, 2023. [https://doi.org/10.1109/TMECH.2022.3226347](https://doi.org/10.1109/TMECH.2022.3226347).
  • [9] CANAL, Rafael. Machine learning applied in automotive ECUs for real-time engine behavior analysis. 2023. Dissertação (Mestrado em Engenharia de Sistemas Eletrônicos) – Universidade Federal de Santa Catarina, Campus Joinville, Joinville, 2023.
  • [10] COELHO, Pedro Miguel Nogueira. Rumo à Indústria 4.0. 2016. Estudo Geral, v. 20, p. c3, 2016. Disponível em: [https://estudogeral.uc.pt/bitstream/10316/36992/1/Tese%20Pedro%20Coelho%20Rumo](https://estudogeral.uc.pt/bitstream/10316/36992/1/Tese%20Pedro%20Coelho%20Rumo). Acesso em: 2 maio 2025.
  • [11] GARG, Prasoon; SILVAS, Emilia; WILLEMS, Frank P. T. Potential of machine learning methods for robust performance and efficient engine control development. IFAC-PapersOnLine, v. 54, n. 10, p. 189–195, 2021. DOI: 10.1016/j.ifacol.2021.10.162.
  • [12] HOCHREITER, S.; SCHMIDHUBER, J. Long short-term memory. Neural Computation, v. 9, n. 8, p. 1735–1780, 1997. DOI: 10.1162/neco.1997.9.8.1735.
  • [13] GOMAR, Shaghayegh; MIRHASSANI, Mitra; AHMADI, Majid. Precise digital implementations of hyperbolic tanh and sigmoid function. In: Asilomar Conference on Signals, Systems and Computers, 2016. IEEE, 2016. p. 1586–1589. DOI: 10.1109/ACSSC.2016.7869625.
  • [14] IAKSCH, Jaqueline Sebastiany. Modelo sistêmico para tratamento de dados no contexto da manutenção de equipamentos agrícolas em uma usina de cana-de-açúcar como apoio à implantação de sistemas de agricultura 4.0. Curitiba: Universidade Tecnológica Federal do Paraná, 2023.
  • [15] WU, Y. et al. Intelligent systems with applications. Intelligent Systems with Applications, v. 22, art. 200395, 2024. DOI: 10.1016/j.iswa.2024.200395.
  • [16] JENSEN, Michael C. The modern industrial revolution, exit, and the failure of internal control systems. The Journal of Finance, v. 48, n. 3, p. 831–880, jul. 1993. Disponível em: [https://onlinelibrary.wiley.com/doi/10.1111/j.1540-6261.1993.tb04022.x](https://onlinelibrary.wiley.com/doi/10.1111/j.1540-6261.1993.tb04022.x).
  • [17] KWON, Daeil; HODKIEWICZ, Melinda R.; FAN, Jiajie; SHIBUTANI, Tadahiro; PECHT, Michael G. IoT-based prognostics and systems health management for industrial applications. IEEE Access, v. 4, p. 3659–3670, 2016. [https://doi.org/10.1109/ACCESS.2016.2587754](https://doi.org/10.1109/ACCESS.2016.2587754).
  • [18] LATINOV, M., FIORINI, N., VICHI, G., INNOCENTI, A. et al. A Machine Learning Approach for Engine Model-Based Control on NOx Emissions. SAE Int. J. Adv. & Curr. Prac. in Mobility 6(3):1574-1582, 2024. [https://doi.org/10.4271/2023-32-0157](https://doi.org/10.4271/2023-32-0157).
  • [19] LIN, Jyong; CHEN, Shih-Chang; SHIH, Yu-Tsen; CHEN, Shi-Huang. A study on remote on-line diagnostic system for vehicles by integrating the technology of OBD, GPS, and 3G. International Journal of Humanities and Social Sciences, v. 3, n. 8, p. 1705–1711, 2009.
  • [20] LÜTTENBERG, Hedda; BARTELHEIMER, Christian; BEVERUNGEN, Daniel. Designing predictive maintenance for agricultural machines. In: Proceedings of the 26th European Conference on Information Systems (ECIS 2018), Portsmouth, UK, 23–28 jun. 2018.
  • [21] LUTCHEN, Ralf; KRÄTSCHMER, Andreas; REUSS, Hans Christian. AI-based classification of CAN measurements for network and ECU identification. Automotive and Engine Technology, v. 7, p. 317–330, 2022. DOI: 10.1007/s41104-022-00116-6.
  • [22] NADJ, Mario; JEGADEESAN, Harshavardhan; MAEDCHE, Alexander; HOFFMANN, Dirk; ERDMANN, Philipp. A situation awareness driven design for predictive maintenance systems: the case of oil and gas pipeline operations. In: EUROPEAN CONFERENCE ON INFORMATION SYSTEMS (ECIS), 24., 2016, Istanbul, Turkey.
  • [23] ORTIZ, Jesús Hamilton. Industry 4.0. Rijeka: IntechOpen, 2020. 304 p. DOI: 10.5772/intechopen.86000.
  • [24] PECHT, M. Prognostics and health management of electronics. New York: Wiley-Interscience, 2009.
  • [25] POÓR, Peter; ŽENÍŠEK, David; BASL, Josef. Historical overview of maintenance management strategies: development from breakdown maintenance to predictive maintenance in accordance with four industrial revolutions. In: INTERNATIONAL CONFERENCE ON INDUSTRIAL ENGINEERING AND OPERATIONS MANAGEMENT, 2019, Pilsen, Czech Republic.
  • [26] REIF, Konrad. Automotive Mechatronics: Automotive Networking, Driving Stability Systems, Electronics. 2nd ed. Wiesbaden: Springer Vieweg, 2015.
  • [27] SALKIN, Ceren; ONER, Mahir; USTUNDAG, Alp; CEVIKCAN, Emre. A conceptual framework for Industry 4.0. In: USTUNDAG, Alp; CEVIKCAN, Emre (Ed.). Industry 4.0: managing the digital transformation. Cham: Springer International Publishing, 2018. p. 3–23. DOI: 10.1007/978-3-319-57870-5\_1.
  • [28] SERGI, B. S.; POPKOVA, E. G.; BOGOVIZ, A. V.; LITVINOVA, T. N. Understanding Industry 4.0: AI, the Internet of Things, and the Future of Work. Bingley: Emerald Publishing Limited, 2019. ISBN 9781789733136.
  • [29] SPEAR, Brian. James Watt: the steam engine and the commercialization of patents. World Patent Information, v. 30, n. 1, p. 53–58, 2008. DOI: [https://doi.org/10.1016/j.wpi.2007.05.009](https://doi.org/10.1016/j.wpi.2007.05.009).
  • [30] WANG, Haitao; YANG, Jie; WANG, Ruihua; SHI, Lichen. Remaining useful life prediction of bearings based on convolution attention mechanism and temporal convolution network. IEEE Access, v. 11, p. 24407–24419, 2023. DOI: 10.1109/ACCESS.2023.3255891.
  • [31] WOLFERT, Sjaak; GE, Lan; VERDOUW, Cor; BOGAARDT, Marc-Jeroen. Big Data in Smart Farming – a review. Agricultural Systems, v. 153, p. 69–80, 2017. DOI: 10.1016/j.agsy.2017.01.023.
  • [32] XU, Tiantian; HAN, Guangjie; ZHU, Hongbo; TALEB, Tarik; PENG, Jinlin. Multi-resolution LSTM-based prediction model for remaining useful life of aero-engine. IEEE Transactions on Vehicular Technology, v. 73, n. 2, p. 1931–1941, 2024. DOI: 10.1109/TVT.2023.3319377.
  • [33] ZHOLONDKOVSKIY, Vladislav V.; LANDYSHEV, Sergey; BOCHAROV, Yuri; BUTUZOV, Vladimir. LSTM-type neural network implementation on a processor based on Neuromatrix and RISC cores for resource-limited applications. In: INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING AND ITS APPLICATIONS (DSPA), 22., 2020. IEEE, 2020. DOI: 10.1109/DSPA48919.2020.9213291.
Como citar:

FERNANDES, Thiago Bastos; PONTE, Leonardo Felipe Nunes; WATERMANN, Matheus Gomes Camargo; BORSATO, Milton; "Abordagem Conceitual para Implementação de Redes Neurais LSTM em C para Prognóstico e Gerenciamento de Saúde de Veículos", p-77-85. In: Anais do XXXII Simpósio Internacional de Engenharia. São Paulo: Blucher, 2025.
ISSN 23577592, DOI 10.5151/simea2025-PAP18

últimos 30 dias

69
downloads

81
visualizações

157
indexações