Prediction of gait trajectories based on the Long Short Term Memory neural networks
Open accessThe DOI matches its Crossref record; 103 of the 103 references it cites were found in a registry or an index.
Abstract
The forecasting of lower limb trajectories can improve the operation of assistive devices and minimise the risk of tripping and balance loss. The aim of this work was to examine four Long Short Term Memory (LSTM) neural network architectures (Vanilla, Stacked, Bidirectional and Autoencoders) in predicting the future trajectories of lower limb kinematics, i.e. Angular Velocity (AV) and Linear Acceleration (LA). Kinematics data of foot, shank and thigh (LA and AV) were collected from 13 male and 3 female participants (28 ± 4 years old, 1.72 ± 0.07 m in height, 66 ± 10 kg in mass) who walked for 10 minutes at preferred walking speed (4.34 ± 0.43 km.h -1 ) and at an imposed speed (5km.h -1 , 15.4% ± 7.6% faster) on a 0% gradient treadmill. The sliding window technique was adopted for training and testing the LSTM models with total kinematics time-series data of 10,500 strides. Results based on leave-one-out cross validation, suggested that the LSTM autoencoders is the top predictor of the lower limb kinematics trajectories (i.e. up to 0.1s). The normalised mean squared error was evaluated on trajectory predictions at each time-step and it obtained 2.82–5.31% for the LSTM autoencoders. The ability to predict future lower limb motions may have a wide range of applications including the design and control of bionics allowing improved human-machine interface and mitigating the risk of falls and balance loss.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“| LSTM | Resolve long-term dependencies through gated | Dynamically parse continuous motion | Exoskeleton joint angle prediction 148,244,245 |”
From Recent advances in intelligent wearable systems: from multiscale biomechanical features towards human motion intent prediction (2026) · cites this paper as [148]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Learning representations by back-propagating errors
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986) · Nature · doi:10.1038/323533a0 · Cited by 31,892 · cited 1×
“Proposed by Rumelhart et al ,. 1986 [63], the SGD algo-rithm aims to obtain the minimum error (MAE in this work) at each batch using the network weights and biases.”
Method(2016) · Cited by 11,679
Biomechanics and motor control of human movement
(1990) · Choice Reviews Online · doi:10.5860/choice.28-2135 · Cited by 8,451
Biomechanics and Motor Control of Human Movement
Winter, D. A. (2009) · Wiley eBooks · doi:10.1002/9780470549148 · Cited by 6,066 · cited 1×
“Additionally, a known 0.1s future gait trajectory falls in the range of slow and fast twitch motor units [73] and may allow a response time (10-120ms) for an elderly patient to adjust their gait and avoid an imminent risk of tripping or balance loss [73–77].”
BackgroundFramewise phoneme classification with bidirectional LSTM and other neural network architectures
Graves, A., & Schmidhuber, J. (2005) · Neural Networks · doi:10.1016/j.neunet.2005.06.042 · Cited by 5,839
Large-Scale Machine Learning with Stochastic Gradient Descent
Bottou, L. (2010) · Proceedings of COMPSTAT'2010 · doi:10.1007/978-3-7908-2604-3_16 · Cited by 5,795
Greedy layer-wise training of deep networks
Bengio, Y., Lamblin, P., Popovici, D., & Larochelle, H. (2007) · Advances in Neural Information Processing Systems 19 · doi:10.7551/mitpress/7503.003.0024 · Cited by 4,841
(2013) · Cited by 3,740
Supervised Sequence Labelling with Recurrent Neural Networks
Graves, A. (2012) · Studies in Computational Intelligence · doi:10.1007/978-3-642-24797-2 · Cited by 3,120
Research works printed but not linked: 4
- Mitchell TM. Machine learning. McGraw-hill New York; 1997.
- Graves A. Generating sequences with recurrent neural networks. arXiv preprint arXiv:13080850. 2013.
- Butterworth S. On the theory of filter amplifiers. Wireless Engineer. 1930; 7(6): 536– 41.
- Soutas-Little RW. Motion analysis and biomechanics. J Rehabil Res Dev. 1998: 49– 68.
Cited by
Deep Learning Approaches for Enhanced Lower-Limb Exoskeleton Control: A Review
Belal, M., Alsheikh, N., Aljarah, A., & Hussain, I. (2024) · IEEE Access · doi:10.1109/access.2024.3414175 · Cited by 42
Kolaghassi, R., Marcelli, G., & Sirlantzis, K. (2023) · IEEE Access · doi:10.1109/access.2023.3252916 · Cited by 42
“capture systems [19], [27], [28], [29], IMUs [26], [30],”
Gait Trajectory Prediction on an Embedded Microcontroller Using Deep Learning
Karakish, M., Fouz, M. A., & ELsawaf, A. (2022) · Sensors · doi:10.3390/s22218441 · Cited by 30
“[38] used different LSTM neural network architectures (Vanilla, Stacked, Bidirectional and Autoencoders) in predicting lower limb kinematic parameters (angular velocity and linear acceleration of the thigh, shank and foot) up to 100 ms (five time frames) ahead, and they reported that the best result was achieved using…”
MethodSensors and algorithms for locomotion intention detection of lower limb exoskeletons
Wang, D., Gu, X., & Yu, H. (2023) · Medical Engineering & Physics · doi:10.1016/j.medengphy.2023.103960 · Cited by 31
BlazePose-Seq2Seq: Leveraging Regular RGB Cameras for Robust Gait Assessment
Hulleck, A. A., Shehhi, A. A., Rich, M. E., Khan, R., Katmah, R., Mohseni, M., Arjmand, N., & Khalaf, K. A. (2024) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2024.3391908 · Cited by 29
Kian, A., Widanapathirana, G., Joseph, A. M., Lai, D. T. H., & Begg, R. K. (2022) · Sensors · doi:10.3390/s22062244 · Cited by 31
Self-citationDing, G., Plummer, A. R., & Georgilas, I. P. (2022) · Frontiers in Bioengineering and Biotechnology · doi:10.3389/fbioe.2022.1021505 · Cited by 24
“The NRMSE value, which has been receiving widespread attention recently, quanti fi es the model error in normalized form (Ardestani et al., 2014; Giarmatzis et al., 2020; Zaroug et al., 2021).”
BackgroundInfluentialAI-Driven Arm Movement Estimation for Sustainable Wearable Systems in Industry 4.0
Muntean, E., Leba, M., & Ionica, A. C. (2025) · Sustainability · doi:10.3390/su17146372 · Cited by 21
Human heart health prediction using GAIT parameters and machine learning model
Singh, P. K., Kourav, P. S., Mohapatra, S., Kumar, V., & Panda, S. K. (2023) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2023.105696 · Cited by 21
Kolaghassi, R., Al-Hares, M. K., Marcelli, G., & Sirlantzis, K. (2022) · Sensors · doi:10.3390/s22082969 · Cited by 24
“[29] also compared between vanilla, stacked, bidirectional, and autoencoder LSTMs while Zhu et al.”
BackgroundChae, S., Choi, A., Jung, H., Kim, T. H., Kim, K.-R., & Mun, J. H. (2021) · Applied Sciences · doi:10.3390/app112411735 · Cited by 12
Using Deep Learning to Predict Minimum Foot–Ground Clearance Event from Toe-Off Kinematics
Asogwa, C. O., Nagano, H., Wang, K., & Begg, R. K. (2022) · Sensors · doi:10.3390/s22186960 · Cited by 12
“These sequences are periodic events that can be converted into digital time-series representations suitable for human gait prediction [24,30].”
BackgroundMethodSelf-citationChen, S., Peng, C., Yang, B., Lin, J., Zhou, L., Jiang, Z., Liu, Z., Liu, Y., & Tang, L. (2026) · npj Artificial Intelligence · doi:10.1038/s44387-026-00083-5 · Cited by 9
Yang, C., Jin, P., & Chen, Y. (2025) · Scientific Reports · doi:10.1038/s41598-025-88284-x · Cited by 9
Lv, X., Tao, Y., & Xue, Y. (2025) · Applied Sciences · doi:10.3390/app15168871 · Cited by 9
Active Gait Retraining with Lower Limb Exoskeleton Based on Robust Force Control
Rosales-Luengas, Y., Salazar, S., Rangel-Popoca, S. J., Cortés-García, Y., Flores, J., & Lozano, R. (2025) · Applied Sciences · doi:10.3390/app15074032 · Cited by 8
Yu, Z., Zhao, J., Chen, D., Chen, S., & Wang, X. (2023) · Journal of Intelligent & Robotic Systems · doi:10.1007/s10846-023-01963-7 · Cited by 8
Saoud, L. S., & Hussain, I. (2023) · IEEE-RAS International Conference on Humanoid Robots (Humanoids) · doi:10.1109/humanoids57100.2023.10375196 · Cited by 8
“Existing literature on deep learning models for gait prediction has primarily focused on short-term predictions with limited output window sizes [24], [25], [26], [27].”
BackgroundEstimation of the legs’ state of a mobile robot based on Long Short-Term Memory network
Albadin, A., Albitar, C., & Saba, M. A. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109539 · Cited by 7
Fleischmann, S., Dietz, S., Shanbhag, J., Wuensch, A., Nitschke, M., Miehling, J., Wartzack, S. J., Leyendecker, S., Eskofier, B. M., & Koelewijn, A. D. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3702646 · Cited by 7
“For example, machine learning models have been applied to characterize movement patterns [7], classify pathological gait [10], assess the risk of falls or injuries [31, 56], predict human motion [62] or gain insights into the control of movement [12].”
Implementing Gait Kinematic Trajectory Forecasting Models on an Embedded System
Shayne, M., Molina, L. A., Hu, B., & Chomiak, T. (2024) · Sensors · doi:10.3390/s24082649 · Cited by 5
Interpretable long-term gait trajectory prediction based on Interpretable-Concatenation former
Yin, J., Chen, M., Zhang, C., Xue, T., Zhang, M., & Zhang, T. (2023) · Robotica · doi:10.1017/s0263574723001492 · Cited by 4
“As an important method for processing sequence information, the long short-term memory (LSTM) model is widely used to predict gait trajectory [5, 6, 7, 8, 9].”
MethodImproving Knee Joint Angle Prediction Through Dynamic Contextual Focus and Gated Linear UnitsPreprint
Saoud, L. S., Ibrahim, H., Aljarah, A., & Hussain, I. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4591943 · Cited by 3
Pitto, L., Simon, F. R., Ertel, G. N., Gauchard, G. C., & Mornieux, G. (2025) · Sensors · doi:10.3390/s25010279 · Cited by 2
Yamashita, J. V. Y. B., Leite, J. P., & Machado, J. B. (2025) · Technologies · doi:10.3390/technologies13110497 · Cited by 3
“For predictive modeling, LSTM networks effectively capture temporal dependencies in gait data, showing good predictive accuracy in terms of Root Mean Squared Error (RMSE) and other metrics [29,30].”
Hosseini, S. N., Mojibi, A., Mohseni, M., Arjmand, N., & Taheri, A. (2026) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2026.3697329 · Cited by 2
“For instance, long short-term memory (LSTM) models have been used to predict gait trajectory [27], [28] and a comparative analysis of LSTM, FCN, convolutional neural networks (CNNs), and trans-former deep-learning models was conducted for predicting gait trajectories [29].”
InfluentialTime Series Classification for Predicting Biped Robot Step Viability
Igual, J., Parik-Americano, P., Becman, E. C., & Forner‐Cordero, A. (2024) · Sensors · doi:10.3390/s24227107 · Cited by 3
Mo, J., Xiong, Q., Chen, Y., Liu, Y., Wu, X., Xiao, N., & Hou, W. (2025) · BioMedical Engineering OnLine · doi:10.1186/s12938-025-01360-1 · Cited by 2
“The key advantage of LSTM is its ability to account for the order of values in input sequences, enabling it to learn long-term dependencies [23].”
InfluentialData-driven Gait-predictive Model for Anticipatory Prosthesis Control
Dey, S., & Schilling, A. F. (2022) · International Conference on Rehabilitation Robotics · doi:10.1109/icorr55369.2022.9896505 · Cited by 4
“thereby predict the desired locomotion at a future time point to minimize latency [14], [15].”
BackgroundPrediction of Gait Trajectory for Motion Mode Switching of Exoskeleton Robots Based on Deep Learning
Tian, X., Liu, R., Lu, Z., Ren, X., Du, Y., Liang, B., & Liu, D. (2024) · IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems · doi:10.1109/cyber63482.2024.10749641 · Cited by 1
Mileti, I., Taborri, J., Pasinetti, S., Lancini, M., Rossi, S., & Patanè, F. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3568533 · Cited by 1
Improving knee joint angle prediction through Dynamic Contextual Focus and Gated Linear Units
Saoud, L. S., Ibrahim, H., Aljarah, A., & Hussain, I. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.111119 · Cited by 2
Data-Driven Estimation of Rowing Forces and Power via Long Short-Term Memory Neural Networks
Meyer, L., Schmid, A., & Braun, S. (2024) · Bulletin of pioneering researches of medical and clinical science · doi:10.51847/gjr2udtuzl · Cited by 1
Predicting Sagittal-Plane Swing Hip Kinematics in Response to Trips
Danforth, S. M., Liu, X., Ward, M. J., Holmes, P., & Vasudevan, R. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3184014 · Cited by 1
Predicting Sagittal-Plane Swing Hip Kinematics in Response to Trips
Danforth, S. M., Liu, X., Ward, M. J., Holmes, P., & Vasudevan, R. (2022) · IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob) · doi:10.1109/biorob52689.2022.9925354 · Cited by 1
“Previous studies that investigate predictive models for human kinematics report results for only one class of model [12]; use simulated trip data instead of human subject experimental data [13]; or use unperturbed data, where less variation is likely present [11].”
BackgroundMethodGait Step Length Classification Using Force Myography
Khatavkar, R., Tiwari, A. K., Bajpai, R., & Joshi, D. (2022) · International Conference for Advancement in Technology (ICONAT) · doi:10.1109/iconat53423.2022.9726014
Realtime Gait Phase Prediction and Early Detection of Gait Abnormalities
Jha, C., Jha, S. K., Bandyopadhyay, O., & Nandi, D. (2024) · International Conference on Computer, Communication, Control and Information Technology · doi:10.1109/c3it60531.2024.10829427
“Four different Long Short-Term Memory (LSTM) architecture types were investigated in [3] for the purpose of forecasting lower limb kinematics.”
Civil Aviation Passenger Throughput Forecasting Model Based on Machine Learning
Yang, L., & Yang, D. (2024) · International Conference on Artificial Intelligence, Deep Learning and Neural Networks (AIDLNN) · doi:10.1109/aidlnn65358.2024.00044
Enhancing Gait Trajectory Prediction through an Adaptive Template Selection Mechanism
Yang, Y., Liu, X., Chen, Y., Shi, J., Liu, Y., Jin, Y., & Wang, X. (2024) · International Conference on Neural Information Processing · doi:10.24135/iconip21
Li, L., Sun, S., Zhu, H., Zheng, C., & Zeng, Y. (2025) · PLoS ONE · doi:10.1371/journal.pone.0323718
Yeoh, J. K. X., Kouzbary, M. A., Osman, N. A. A., Chan, C. K., Shasmin, H. N., & Tham, L. K. (2025) · International Journal of Control · doi:10.1080/00207179.2025.2537253
“Moreover, this system exhibits better performance than the LSTM neural network developed by Zaroug et al. (2021) which achieved an Xcorr of up to 0.96 in predicting gait trajectories.”
InfluentialJiang, J., Saadat, M., Liu, G., Maddalena, M., & Rezaei, S. M. (2025) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2025.3619996
Interlimb Gait Trajectory Synchronization from Inertial Measurement Unit (IMU) with ResUNet
Tantisiriwat, N., Ngamdi, N., Khawkhom, A., Arnin, J., Wongsawat, Y., & Anopas, D. (2025) · Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc58623.2025.11251528
“The future trajectory of the lower limbs was accurately predicted within a 200-millisecond-long window using LSTMs [17], [18].”
Jiang, J., Liu, G., & Saadat, M. (2025) · IEEE International Conference on Cyborg and Bionic Systems (CBS) · doi:10.1109/cbs65871.2025.11267708
“The most selected installation positions of IMU are ankle joint, calf, thigh, instep and whole body, while the number of installations ranges from 1 to 11 [7]-[8].”
Inertial sensor-based step length estimation model
Vezočnik, M. (2023) · Repository of the University of Ljubljana (University of Ljubljana)
Optimizing Data Collection for Bodily Emotion Recognition: A Comparative Study
Cho, Y., Jung, M., Bae, J., & Kim, K. (2025) · IEEE Access · doi:10.1109/access.2025.3634675
Sistema de controle para prótese ativa de joelho utilizando Deep Learning embarcado
Yamashita, J. V. Y. B. (2024) · LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas)
Li, N., Li, L., He, B., & Zhao, J. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.109697
Minimum Foot Clearance Prediction in Stroke Survivors: A Transformer-Based Approach
Sengupta, N., Begg, R. K., Rao, A. S., Desai, N., Bajelan, S., Said, C. M., & Palaniswami, M. (2026) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2026.3694895
“Furthermore, LSTM [22] and di ff erent variants of LSTM models [23] were proposed to predict gait trajectory.”
Self-citationLee, S. J., Choi, J., Ahn, J., & Kang, K. (2026) · ACM Symposium on Applied Computing · doi:10.1145/3748522.3779948
“Zaroug et al. [28] predicted future trajectories of lower limb kinematics, i.e., angular velocity and linear acceleration, using Long Short-Term Memory (LSTM).”
Patani, A., Patel, B., Pandit, R., Bagatharia, S., & Patel, A. (2026) · Frontiers in Sports and Active Living · doi:10.3389/fspor.2026.1906206
Porter-Honicky, N., & Rouse, E. J. (2026) · IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob) · doi:10.1109/biorob66782.2026.11681876
“In recent years, machine learning techniques have been applied to a myriad of challenges in biomechanics, including time series forecasting of gait kinematics and kinetics [16], [17], [18], [19].”
Pansardi, O. (n.d.) · Institutional Research Information System University of Turin (University of Turin)
Comparison of Regression Models for Predicting Lower-Limb Joint Angles During Stationary Cycling
Ahmadi, R., Rasoulian, S., Kakavand, R., Tahghighi, P., Herzog, W., & Komeili, A. (2026) · AI Sensors · doi:10.3390/aisens2040014
Metadata
- Authors
- Zaroug, A., Garofolini, A., Lai, D. T. H., Mudie, K., & Begg, R.
- Affiliations
- Victoria University (AU)
- Defence Science and Technology Group (AU)
- Corresponding author
- Zaroug, Abdelrahman
- Garofolini, Alessandro
- Lai, Daniel T. H.
- Mudie, Kurt
- Begg, Rezaul
- Venue
- PLOS ONE
- Publisher
- Public Library of Science (PLoS)
- Published
- 2021-08-05
- Volume
- 16
- Issue
- 8
- Pages
- e0255597
- ISSN
- 1932-6203 (ISSN-L)
- PMC
- PMC8341582
- OpenAlex
- W3187228760
- Cited by
- 48
- References
- 103
- Citations by year
- 2026: 6 · 2025: 15 · 2024: 10 · 2023: 7 · 2022: 9 · 2021: 1
- FWCI
- 2.05
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS ONE (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Victoria University Research Repository (Victoria University) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Figshare (submittedVersion, cc-by)
- Funders
- Victoria University
- Defence Science and Technology Group
- Topics
- Physical Sciences › Engineering › Biomedical Engineering › Prosthetics and Rehabilitation Robotics
- Health Sciences › Health Professions › Physical Therapy, Sports Therapy and Rehabilitation › Balance, Gait, and Falls Prevention
- Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
- Keywords
- lower limb kinematics, long short-term memory, LSTM autoencoder, angular velocity, linear acceleration, walking speed, treadmill walking, assistive devices, fall prevention, bionic prosthetics, human-machine interaction
- MeSH terms
- Adult, Algorithms, Biomechanical Phenomena, Female, Gait, Humans, Male, Memory, Short-Term, Walking, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.