Evaluation of Lower Limb Gait Pattern Generation from BMI-Based Angular Velocity Estimation for Personalized Robotic Rehabilitation
DOI:
https://doi.org/10.3991/ijoe.v22i10.62264Keywords:
Gait pattern generation, joint angle estimation, angular velocity, Body Mass Index, rehabilitation robot, gait similarity, lower limb biomechanics, personalized gait control, PicobotAbstract
This study assesses the ability of gait angle patterns produced by a phase-averaged velocity estimation NN-15 model based on BMI and walking speed to serve as personalized joint motion references. A total of 424 gait cycles from 24 seen subjects and 131 gait cycles from 8 unseen subjects across four BMI categories (underweight (U), normal (N), overweight (Ov), and obese (Ob)) and both limb sides were analyzed. Similarity was quantified using root mean square error (RMSE), mean absolute error (MAE), Pearson correlation coefficient (r), and Bland–Altman analysis. The hip joint achieved the strongest agreement (r = 0.82–0.94, RMSE = 4.24–8.42°), followed by the ankle (r = 0.19–0.81, RMSE = 6.17–12.39°) and the knee (r = 0.15–0.83, RMSE = 9.28–23.80°). The Ov group consistently demonstrated the best performance at the hip and ankle. Significant left–right accuracy at the hip remained bilaterally accurate. Based on these findings, the NN-15 framework is validated as a personalized hip joint reference for a rehabilitation robot called Picobot. However, adopting different gait phase schemes is advised for the ankle, and a position-hold strategy during Phases 1 and 2 is proposed for the knee to improve joint estimation.
References
[1] T. Thayabaranathan et al., “Global stroke statistics 2022,” Oct. 01, 2022, SAGE Publications Inc. doi: 10.1177/17474930221123175.
[2] A. Kusec et al., “Long-term psychological outcomes following stroke: the OX-CHRONIC study,” BMC Neurol., vol. 23, no. 1, Dec. 2023, doi: 10.1186/s12883-023-03463-5.
[3] R. Algahtani, R. Gasemaltayeb, Z. Binsiddiq, M. Almatrafi, S. Felimban, and R. Alsaud, “Therapeutic interventions for the non-motor strokes sequelae: an overview,” Egyptian Journal of Neurology, Psychiatry and Neurosurgery, vol. 61, no. 1, Dec. 2025, doi: 10.1186/s41983-025-00969-2.
[4] H. Nedergård, M. Sandlund, C. K. Häger, and S. Palmcrantz, “Users’ experiences of intensive robotic-assisted gait training post-stroke–‘a push forward or feeling pushed around?,’” Disabil. Rehabil., vol. 45, no. 23, pp. 3861–3868, 2023, doi: 10.1080/09638288.2022.2140848.
[5] P. Capodaglio et al., “Effect of obesity on knee and ankle biomechanics during walking,” Sensors, vol. 21, no. 21, Nov. 2021, doi: 10.3390/s21217114.
[6] Malek Adouni, Fadi Alkhatib, Raouf Hajji, and Tanvir R. Faisal, “Effects of overweight and obesity on lower limb walking characteristics from joint kinematics to muscle activations,” Gait Posture, vol. 113, pp. 337–344, Sep. 2024, doi: https://doi.org/10.1016/j.gaitpost.2024.06.024.
[7] T. Van Criekinge, W. Saeys, S. Truijen, L. Vereeck, L. H. Sloot, and A. Hallemans, “A full-body motion capture gait dataset of 138 able-bodied adults across the life span and 50 stroke survivors,” Sci. Data, vol. 10, no. 1, Dec. 2023, doi: 10.1038/s41597-023-02767-y.
[8] Rifdatun Ni’mah, Dimas Adiputra, Helisyah Nur Fadhilah, Mohamad Azlan Mohamed Shapie, and Mohd Hafiz Awang Hassim, “Understanding Normative Gait Spatio-Temporal variability: A Multi-Factor Analysis of Age, Gender, and BMI,” Gait Posture, vol. 212, no. ESMAC Abstract 2025, pp. 4–5, Sep. 2025, doi: 10.1016/j.cmpb.2021.106414.
[9] R. R. Rizkika, H. N. Fadhilah, T. Mustaqim, and R. Ni’Mah, “Development of a Document-Based Gait System With Interactive Visualisation for Clinical Analysis,” Jurnal RESTI, vol. 9, no. 3, pp. 554–563, Jun. 2025, doi: 10.29207/resti.v9i3.6451.
[10] V. B. Semwal, R. Jain, P. Maheshwari, and S. Khatwani, “Gait reference trajectory generation at different walking speeds using LSTM and CNN,” Multimed. Tools Appl., vol. 82, no. 21, pp. 33401–33419, Sep. 2023, doi: 10.1007/s11042-023-14733-2.
[11] F. Wang et al., “Estimation of Lower Limb Joint Angles and Joint Moments during Different Locomotive Activities Using the Inertial Measurement Units and a Hybrid Deep Learning Model,” Sensors, vol. 23, no. 22, Nov. 2023, doi: 10.3390/s23229039.
[12] D. Shi, W. Zhang, X. Ding, and L. Sun, “Parametric generation of three-dimensional gait for robot-assisted rehabilitation,” Biol. Open, vol. 9, no. 3, Mar. 2020, doi: 10.1242/bio.047332.
[13] B. Koopman, E. H. F. van Asseldonk, and H. Van der Kooij, “Speed-dependent reference joint trajectory generation for robotic gait support,” J. Biomech., vol. 47, no. 6, pp. 1447–1458, Apr. 2014, doi: 10.1016/j.jbiomech.2014.01.037.
[14] H. Vallery, E. H. F. Van Asseldonk, M. Buss, and H. Van Der Kooij, “Reference trajectory generation for rehabilitation robots: Complementary limb motion estimation,” in IEEE Transactions on Neural Systems and Rehabilitation Engineering, Feb. 2009, pp. 23–30. doi: 10.1109/TNSRE.2008.2008278.
[15] D. Adiputra et al., “Estimation of Lower Limb Joint Angular Velocity from BMI and Walking Speed Using Regression and Neural Network Models for Personalized Gait Control,” IEEE Access, vol. 13, Nov. 2025, doi: 10.1109/ACCESS.2025.3633220.
[16] D. Adiputra, M. A. A. Rahman, Ubaidillah, and S. A. Mazlan, “Improving passive ankle foot orthosis system using estimated ankle velocity reference,” IEEE Access, vol. 8, pp. 194780–194794, 2020, doi: 10.1109/ACCESS.2020.3033852.
[17] D. Kim, C. L. Lewis, and S. V. Gill, “Effects of obesity and foot arch height on gait mechanics: A cross-sectional study,” PLoS One, vol. 16, no. 11 November, Nov. 2021, doi: 10.1371/journal.pone.0260398.
[18] V. Rosso, V. Agostini, R. Takeda, S. Tadano, and L. Gastaldi, “Influence of BMI on gait characteristics of young adults: 3D evaluation using inertial sensors,” Sensors (Switzerland), vol. 19, no. 19, Oct. 2019, doi: 10.3390/s19194221.
[19] L. Vismara et al., “Brain Asymmetry and Its Effects on Gait Strategies in Hemiplegic Patients: New Rehabilitative Conceptions,” Brain Sci., vol. 12, no. 6, Jun. 2022, doi: 10.3390/brainsci12060798.
[20] D. Adiputra, U. Asfari, Ubaidillah, M. A. Abdul Rahman, and A. M. Harun, “Immediate Effect Evaluation of a Robotic Ankle–Foot Orthosis with Customized Algorithm for a Foot Drop Patient: A Quantitative and Qualitative Case Report,” Int. J. Environ. Res. Public Health, vol. 20, no. 4, Feb. 2023, doi: 10.3390/ijerph20043745.
[21] L. Mobbs, V. Fernando, R. D. Fonseka, P. Natarajan, M. Maharaj, and R. J. Mobbs, “Normative Database of Spatiotemporal Gait Metrics Across Age Groups: An Observational Case–Control Study,” Sensors, vol. 25, no. 2, Jan. 2025, doi: 10.3390/s25020581.
[22] P. Griffiths and T. Murrells, “Reliability assessment and approaches to determining agreement between measurements: Classic methods paper,” 2010, Elsevier Ltd. doi: 10.1016/j.ijnurstu.2010.03.004.
[23] F. Moissenet, F. Leboeuf, and S. Armand, “Lower limb sagittal gait kinematics can be predicted based on walking speed, gender, age and BMI,” Sci. Rep., vol. 9, no. 1, Dec. 2019, doi: 10.1038/s41598-019-45397-4.
[24] J. Sung et al., “Prediction of lower extremity multi-joint angles during overground walking by using a single IMU with a low frequency based on an LSTM recurrent neural network,” Sensors, vol. 22, no. 1, Jan. 2022, doi: 10.3390/s22010053.
[25] J. Hong, C. Chun, S. J. Kim, and F. C. Park, “Gaussian Process Trajectory Learning and Synthesis of Individualized Gait Motions,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 27, no. 6, pp. 1236–1245, Jun. 2019, doi: 10.1109/TNSRE.2019.2914095.
[26] Richard Baker, Measuring Walking: A Handbook of Clinical Gait Analysis, First. London: Mac Keith Press, 2013.
[27] M. Patil, M. S. Kulkarni, A. Sinha, and R. R. Ghorpade, “Studies on Biomechanics of Human Ankle Joint, Joint Forces Foot Pressure Assessment During Gait,” Trends Biomater. Artif. Organs, vol. 38, no. 3, pp. 153–158, 2024, doi: 10.5281/zenodo.14826276.
[28] J. Lebleu, T. Gosseye, C. Detrembleur, P. Mahaudens, O. Cartiaux, and M. Penta, “Lower limb kinematics using inertial sensors during locomotion: Accuracy and reproducibility of joint angle calculations with different sensor-to-segment calibrations,” Sensors, vol. 20, no. 3, Feb. 2020, doi: 10.3390/s20030715.
[29] M. Al Borno et al., “OpenSense: An open-source toolbox for inertial-measurement-unit-based measurement of lower extremity kinematics over long durations,” J. Neuroeng. Rehabil., vol. 19, no. 1, Dec. 2022, doi: 10.1186/s12984-022-01001-x.
[30] C. Jiang, Y. Yang, H. Mao, D. Yang, and W. Wang, “Effects of Dynamic IMU-to-Segment Misalignment Error on 3-DOF Knee Angle Estimation in Walking and Running,” Sensors, vol. 22, no. 22, Nov. 2022, doi: 10.3390/s22229009.
[31] X. Ma, Y. Liu, Q. Song, and C. Wang, “Continuous estimation of knee joint angle based on surface electromyography using a long short-term memory neural network and time-advanced feature,” Sensors, vol. 20, no. 17, pp. 1–18, Sep. 2020, doi: 10.3390/s20174966.
[32] E. Bergamini, A. Cereatti, and G. Pavei, “Walking symmetry is speed and index dependent,” Sci. Rep., vol. 14, no. 1, Dec. 2024, doi: 10.1038/s41598-024-69461-w.
[33] W. Forczek and R. Staszkiewicz, “An evaluation of symmetry in the lower limb joints during the able-bodied gait of women and men,” J. Hum. Kinet., vol. 35, no. 1, pp. 47–57, 2012, doi: 10.2478/v10078-012-0078-5.
[34] K. Klaewkasikum, T. Patathong, C. Angsanuntsukh, T. Woratanarat, J. Sanguantrakul, and P. Woratanarat, “The ankle kinematic reference of normal gait pattern in Thai adults,” Front. Surg., vol. 9, Aug. 2022, doi: 10.3389/fsurg.2022.915090.
[35] T. McGrath, R. Fineman, and L. Stirling, “An auto-calibrating knee flexion-extension axis estimator using principal component analysis with inertial sensors,” Sensors (Switzerland), vol. 18, no. 6, Jun. 2018, doi: 10.3390/s18061882.
[36] Z. Zhang, X. Cai, M. Zhang, W. Chen, Y. Chen, and P. Wang, “Real‐Time Gait Intention Recognition for Active Control of Unilateral Knee Exoskeleton,” Appl. Bionics Biomech., vol. 2024, no. 1, Jan. 2024, doi: 10.1155/2024/9426782.
[37] P. G. Arauz et al., “Spine and lower body symmetry during treadmill walking in healthy individuals—In-vivo 3-dimensional kinematic analysis,” PLoS One, vol. 17, no. 10 October, Oct. 2022, doi: 10.1371/journal.pone.0275174.
[38] T. Jiroumaru, M. Wachi, Y. Hyodo, Y. Oka, and T. Fujikawa, “Changes in Iliopsoas Muscle Activity and Hip/Pelvic Kinematics With Variations in Step Length and Cadence at a Fixed Walking Speed,” Cureus, May 2025, doi: 10.7759/cureus.84737.
[39] F. Foroughi, S. Sadeh, and H. Y. Hsiao, “Effect of Vertical Ground Reaction Force Biofeedback on Knee and Hip Neuromechanical Characteristics During Walking in Older Adults,” Applied Sciences (Switzerland), vol. 15, no. 6, Mar. 2025, doi: 10.3390/app15062945.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Dimas Adiputra, Helisyah Nur Fadhilah, Rifdatun Ni'mah, Kadek Heri Sanjaya, Kok Swee Sim, Nur Khozin
This work is licensed under a Creative Commons Attribution 4.0 International License.
