A TINYML-AUGMENTED INERTIAL NAVIGATION SYSTEM FOR REAL-TIME DRIFT COMPENSATION ON AN STM32 MICROCONTROLLER
DOI: 
SHARE THIS ARTICLE
DOWNLOAD ARTICLE
REVIEW REPORT
CITE THIS ARTICLE
Low-cost MEMS-based inertial navigation systems (INS) suffer from nonlinear and time-varying gyroscope bias drift, leading to cumulative orientation errors in long-duration applications. Traditional sensor fusion algorithms assume constant bias and do not compensate dynamic drift behavior under operating conditions. This paper presents a TinyML-augmented inertial navigation system implemented on an STM32 microcontroller for real-time adaptive drift compensation. A lightweight neural network model is trained using temporal gyroscope features and deployed using TensorFlow Lite Micro with 8-bit quantization. The estimated bias was removed prior to quaternion-based Madgwick sensor fusion. Experimental validation shows reduced cumulative drift, improved yaw stability, and real-time execution feasibility within strict embedded memory constraints. The proposed approach confirms the integration of embedded machine learning in aerospace navigation systems.
Keywords
Article Information
Article Metrics
HOW TO CITE
Copy
P.S.Karthikkumar, , S.Vignesh, , A, E., A, R. R. & G, N. (2026). A Tinyml-Augmented Inertial Navigation System for Real-Time Drift Compensation on an STM32 Microcontroller. International Journal of Science, Strategic Management and Technology, 02(05). https://doi.org/10.55041/ijsmt.v2i5.134
P.S.Karthikkumar, , et al.. "A Tinyml-Augmented Inertial Navigation System for Real-Time Drift Compensation on an STM32 Microcontroller." International Journal of Science, Strategic Management and Technology, vol. 02, no. 05, 2026, pp. . doi:https://doi.org/10.55041/ijsmt.v2i5.134.
P.S.Karthikkumar, , S.Vignesh,Ezhumalai A,Raghul A, and Nithishwaran G. "A Tinyml-Augmented Inertial Navigation System for Real-Time Drift Compensation on an STM32 Microcontroller." International Journal of Science, Strategic Management and Technology 02, no. 05 (2026). https://doi.org/https://doi.org/10.55041/ijsmt.v2i5.134.
References
[2] R. Mahony, T. Hamel, and J. Pflimlin, “Nonlinear complementary filters on the special orthogonal group,” IEEE Trans. Autom. Control, vol. 53, no. 5, pp. 1203–1218, 2008.
[3] E. Foxlin, “Inertial head-tracker sensor fusion by a complementary separate-bias Kalman filter,” Proc. IEEE VRAIS, pp. 185–194, 1996.
[4] Y. S. Suh, “Orientation estimation using a quaternion-based indirect Kalman filter,” IEEE Trans. Instrum. Meas., vol. 59, no. 12, pp. 3299–3305, 2010.
[5] H. Fourati, “Heterogeneous data fusion algorithm for pedestrian navigation via foot-mounted inertial measurement unit and complementary filter,” IEEE Trans. Instrum. Meas., vol. 64, no. 1, pp. 221–229, 2015.
[6] R. Cechowicz, “Bias drift estimation for MEMS gyroscope used in inertial navigation,” Acta Mech. Autom., vol. 11, no. 2, pp. 104–108, 2017. (researchgate.net)
[7] S. Han et al., “Random error reduction algorithms for MEMS inertial sensors: A review,” Micromachines, vol. 11, no. 11, p. 1021, 2020. (mdpi.com)
[8] S. Han, “Startup drift compensation of MEMS INS based on PSO-GRNN,” Sensors, 2025. (pmc.ncbi.nlm.nih.gov)
[9] M. McManus, “Inertial navigation system drift reduction using scientific machine learning,” M.Eng. thesis, MIT, 2024. (dspace.mit.edu)
[10] C. Chao and J. Zhao, “TinyGC-Net: An extremely tiny network for calibrating MEMS gyroscopes,” arXiv preprint arXiv:2403.02618, 2024. (arxiv.org)
Ethics and Compliance
Indexed In
Similar Articles
Previous Article
