AI-Based Pendulum Motion Recognition and Period Prediction Using Smartphone Sensor Data
Abstract
were extracted from three specific-force channels and one angular-velocity channel. Four lightweight model families were evaluated using
a single stratified 80/20 train-test split. On 260 classification test sequences, random forest achieved 99.62% accuracy and a macro-F1 score
of 99.62%. On 208 non-stationary sequences, support vector regression achieved an RMSE of 0.052 s and R² of 0.962, while random-forest
regression produced an MAE of 0.026 s and MAPE of 1.41%. The results establish a quantitative benchmark for the proposed framework under the defined study conditions. Repeated leakage-controlled validation and the release of code, data, split indices, and final hyperparameters
would further strengthen reproducibility and generalizability.
Keywords
Full Text:
PDFReferences
[1] G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, and L. Yang, "Physics-informed machine learning, " Nature Reviews
Physics, vol. 3, pp. 422-440, 2021. doi: 10.1038/s42254-021-00314-5
[2] S. Cuomo, V. Schiano Di Cola, F. Giampaolo, G. Rozza, M. Raissi, and F. Piccialli, "Scientific machine learning through physicsinformed neural networks: Where we are and what's next, " Journal of Scientifi c Computing, vol. 92, art. 88, 2022. doi: 10.1007/s10915-
022-01939-z
[3] N. Botteghi, M. Guo, and C. Brune, "Deep kernel learning of dynamical models from high-dimensional noisy data, " Scientifi c Reports,
vol. 12, art. 21530, 2022. doi: 10.1038/s41598-022-25362-4
[4] L. Tong, H. Ma, Q. Lin, J. He, and L. Peng, "A novel deep learning Bi-GRU-I model for real-time human activity recognition using inertial sensors, " IEEE Sensors Journal, vol. 22, no. 6, pp. 6164-6174, 2022. doi: 10.1109/JSEN.2022.3148431
[5] D. Garcia-Gonzalez, D. Rivero, E. Fernandez-Blanco, and M. R. Luaces, "New machine learning approaches for real-life human
activity recognition using smartphone sensor-based data, " Knowledge-Based Systems, vol. 262, art. 110260, 2023. doi: 10.1016/
j.knosys.2023.110260
[6] V. E. Di Bacco and W. H. Gage, "Evaluation of a smartphone accelerometer system for measuring nonlinear dynamics during treadmill walking: Concurrent validity and test-retest reliability, " Journal of Biomechanics, vol. 151, art. 111527, 2023. doi: 10.1016/
j.jbiomech.2023.111527
[7] T. Liu and H. Meidani, "Physics-informed neural networks for system identification of structural systems with a multiphysics damping
model, " Journal of Engineering Mechanics, vol. 149, no. 10, art. 04023079, 2023. doi: 10.1061/JENMDT.EMENG-7060
[8] F. Hussain, N. J. Goncalves, D. Alexandre, et al., "A smartphone accelerometer data-driven approach to recognize activities of daily life:
A comparative study, " Smart Health, vol. 30, art. 100432, 2023. doi: 10.1016/j.smhl.2023.100432
[9] J. Ge, G. Xu, J. Lu, X. Xu, L. Li, and X. Meng, "SensorNet: An adaptive attention convolutional neural network for sensor feature learning, " Sensors, vol. 24, no. 11, art. 3274, 2024. doi: 10.3390/s24113274
[10] N. M. Foumani, L. Miller, C. W. Tan, G. I. Webb, G. Forestier, and M. Salehi, "Deep learning for time series classification and extrinsic
regression: A current survey, " ACM Computing Surveys, vol. 56, no. 9, art. 217, 2024. doi: 10.1145/3649448
[11] A. Belendez, A. Hernandez, A. Marquez, T. Belendez, and C. Neipp, "Analytical approximations for the period of a nonlinear pendulum,
" European Journal of Physics, vol. 27, no. 3, pp. 539-551, 2006. doi: 10.1088/0143-0807/27/3/008
[12] T. Chatchawaltheerat, S. Khemmani, and C. Puttharugsa, "Investigating the large angle of a physical pendulum using a smartphone's
sensors, " Physics Education, vol. 56, no. 4, art. 045023, 2021. doi: 10.1088/1361-6552/ac015c
DOI: http://dx.doi.org/10.70711/frim.v4i8.10021
Refbacks
- There are currently no refbacks.
