{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,13]],"date-time":"2026-09-13T10:56:15Z","timestamp":1789296975365,"version":"build-2803163510"},"reference-count":54,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T00:00:00Z","timestamp":1784678400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurorobot."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Accurate gait recognition using wearable sensors is of significant clinical value for adaptive prosthetic control, lower-limb exoskeleton assistance, and objective rehabilitation assessment. However, subject-independent recognition remains a major challenge, as unseen individuals can exhibit highly variable limb kinematics and muscle activation patterns, and existing approaches often rely on a single sensor modality or naive fusion strategies that fail to leverage the complementary information between inertial and electromyographic signals.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>To address these gaps, this study proposes DSAF, a dual-stream attention fusion network that separately encodes kinematic (IMU) and neuromuscular (EMG) information, and adaptively integrates them through a physiological complementarity weighting mechanism designed for window-level modality adaptation. The framework is evaluated on the public HuGaDB dataset for eight common locomotion activities (e.g., walking, running, stair negotiation, and sit-to-stand transitions), using a leave-one-subject-out protocol to rigorously assess generalization to new users.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>DSAF achieves 96.41% accuracy, 96.08% macro-precision, 95.62% macro-recall, and 95.81% macro-F1, consistently outperforming recent sequence-learning baselines across all 18 held-out subjects. Ablation studies further confirm that both the modality-specific encoding and the adaptive fusion mechanism contribute positively to the performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>These findings indicate that adaptive IMU-EMG fusion can effectively strengthen wearable gait recognition, providing a promising solution for real-world rehabilitation monitoring and assistive human-machine interfaces.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fnbot.2026.1901746","type":"journal-article","created":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T05:52:15Z","timestamp":1784699535000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Subject-independent gait activity recognition using DSAF: dual-stream IMU-EMG attention fusion with asymmetric temporal encoding and physiological complementarity weighting"],"prefix":"10.3389","volume":"20","author":[{"given":"Zhangyue","family":"Xu","sequence":"first","affiliation":[{"name":"School of Mechanical and Power Engineering, East China University of Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yan","family":"Zhu","sequence":"additional","affiliation":[{"name":"School of Mathematics, East China University of Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shurui","family":"Li","sequence":"additional","affiliation":[{"name":"School of Mathematics, East China University of Science and Technology","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,22]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"84","DOI":"10.3390\/make7030084","article-title":"A comparison between unimodal and multimodal segmentation models for deep brain structures from T1- and T2-weighted MRI","volume":"7","author":"Altini","year":"2025","journal-title":"Mach. Learn. Knowl. Extr."},{"key":"ref2","article-title":"An empirical evaluation of generic convolutional and recurrent networks for sequence modeling","author":"Bai","year":"2018","journal-title":"arXiv"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1186\/s12984-021-00906-3","article-title":"Review of control strategies for lower-limb exoskeletons to assist gait","volume":"18","author":"Baud","year":"2021","journal-title":"J. Neuroeng. Rehabil."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"143883","DOI":"10.1109\/ACCESS.2024.3414175","article-title":"Deep learning approaches for enhanced lower-limb exoskeleton control: a review","volume":"12","author":"Belal","year":"2024","journal-title":"IEEE Access"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"6084","DOI":"10.3390\/app14146084","article-title":"Enhancing survival analysis model selection through XAI(t) in healthcare","volume":"14","author":"Berloco","year":"2024","journal-title":"Appl. Sci."},{"key":"ref6","doi-asserted-by":"publisher","first-page":"8553","DOI":"10.1109\/JIOT.2019.2920283","article-title":"IoT wearable sensor and deep learning: an integrated approach for personalized human activity recognition in a smart home environment","volume":"6","author":"Bianchi","year":"2019","journal-title":"IEEE Internet Things J."},{"key":"ref7","doi-asserted-by":"publisher","first-page":"6455","DOI":"10.3390\/app12136455","article-title":"Inline defective laser weld identification by processing thermal image sequences with machine and deep learning techniques","volume":"12","author":"Buongiorno","year":"2022","journal-title":"Appl. Sci."},{"key":"ref8","author":"Casale","year":"2011"},{"key":"ref9","first-page":"131","volume-title":"Lecture Notes in Computer Science","author":"Chereshnev","year":"2018"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"645","DOI":"10.1007\/s00779-010-0293-9","article-title":"Preprocessing techniques for context recognition from accelerometer data","volume":"14","author":"Figo","year":"2010","journal-title":"Pers. Ubiquit. Comput."},{"key":"ref12","doi-asserted-by":"publisher","first-page":"1947","DOI":"10.1109\/TNSRE.2026.3682388","article-title":"Muscle network of Parkinson's gait: a 12-month longitudinal analysis before and after deep brain stimulation surgery","volume":"34","author":"Ghislieri","year":"2026","journal-title":"IEEE Trans. Neural Syst. Rehabil. Eng."},{"key":"ref13","doi-asserted-by":"publisher","first-page":"1232","DOI":"10.3390\/prosthesis5040085","article-title":"EMG and IMU data fusion for locomotion mode classification in transtibial amputees","volume":"5","author":"Gonzales-Huisa","year":"2023","journal-title":"PRO"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"5765","DOI":"10.3390\/s25185765","article-title":"Human activity recognition via attention-augmented TCN-BiGRU fusion","volume":"25","author":"He","year":"2025","journal-title":"Sensors"},{"key":"ref15","author":"Hu","year":"2018"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"2866","DOI":"10.3390\/s21082866","article-title":"A lightweight attention-based CNN model for efficient gait recognition with wearable IMU sensors","volume":"21","author":"Huang","year":"2021","journal-title":"Sensors"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"1963","DOI":"10.1109\/JSEN.2023.3338264","article-title":"Smart-wearable sensors and CNN-BiGRU model: a powerful combination for human activity recognition","volume":"24","author":"Imran","year":"2024","journal-title":"IEEE Sensors J."},{"key":"ref18","author":"Jiang","year":"2015"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"20854","DOI":"10.1109\/ACCESS.2024.3357143","article-title":"Human activity recognition via temporal fusion contrastive learning","volume":"12","author":"Kim","year":"2024","journal-title":"IEEE Access"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"506","DOI":"10.1109\/JSEN.2022.3222825","article-title":"Contrastive accelerometer-gyroscope embedding model for human activity recognition","volume":"23","author":"Koo","year":"2023","journal-title":"IEEE Sensors J."},{"key":"ref22","doi-asserted-by":"publisher","first-page":"42335","DOI":"10.1038\/s41598-025-26297-2","article-title":"A tiny inertial transformer for human activity recognition via multimodal knowledge distillation and explainable AI","volume":"15","author":"Lamaakal","year":"2025","journal-title":"Sci. Rep."},{"key":"ref23","doi-asserted-by":"publisher","first-page":"563","DOI":"10.1007\/s11517-008-0327-x","article-title":"Support vector machine for classification of walking conditions using miniature kinematic sensors","volume":"46","author":"Lau","year":"2008","journal-title":"Med. Biol. Eng. Comput."},{"key":"ref24","doi-asserted-by":"publisher","first-page":"9101","DOI":"10.3390\/s23229101","article-title":"Multimodal gait abnormality recognition using a convolutional neural network-bidirectional long short-term memory network based on multi-sensor data fusion","volume":"23","author":"Li","year":"2023","journal-title":"Sensors"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"1711","DOI":"10.1109\/JIOT.2022.3209084","article-title":"Activity-based person identification using multimodal wearable sensor data","volume":"10","author":"Luo","year":"2023","journal-title":"IEEE Internet Things J."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"6291","DOI":"10.3390\/s21186291","article-title":"The role of surface electromyography in data fusion with inertial sensors to enhance locomotion recognition and prediction","volume":"21","author":"Meng","year":"2021","journal-title":"Sensors"},{"key":"ref28","doi-asserted-by":"publisher","first-page":"12592","DOI":"10.3390\/app152312592","article-title":"Attention-based CNN-BiGRU-transformer model for human activity recognition","volume":"15","author":"Miao","year":"2025","journal-title":"Appl. Sci."},{"key":"ref29","doi-asserted-by":"publisher","first-page":"115","DOI":"10.3390\/s16010115","article-title":"Deep convolutional and LSTM recurrent neural networks for multimodal wearable activity recognition","volume":"16","author":"Ordonez","year":"2016","journal-title":"Sensors"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"260","DOI":"10.3390\/s25010260","article-title":"A deep learning-based framework oriented to pathological gait recognition with inertial sensors","volume":"25","author":"Palazzo","year":"","journal-title":"Sensors"},{"key":"ref31","author":"Palazzo","year":""},{"key":"ref32","doi-asserted-by":"publisher","first-page":"2350024","DOI":"10.1142\/S0219519423500240","article-title":"Coherence networks of bilateral upper limb motions in chronic stroke patients","volume":"23","author":"Pan","year":"2023","journal-title":"J. Mech. Med. Biol."},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-026-54191-y","article-title":"A serious game for assessing upper-limb visuomotor adaptation in children with cerebral palsy during reaching tasks in virtual reality","author":"Pasquini","year":"2026","journal-title":"Sci. Rep."},{"key":"ref34","author":"Potluri","year":"2019"},{"key":"ref35","doi-asserted-by":"publisher","first-page":"2727","DOI":"10.3390\/s21082727","article-title":"Wearable sensor-based real-time gait detection: a systematic review","volume":"21","author":"Prasanth","year":"2021","journal-title":"Sensors"},{"key":"ref36","doi-asserted-by":"publisher","first-page":"36","DOI":"10.3390\/diagnostics15010036","article-title":"Validity of wearable inertial sensors for gait analysis: a systematic review","volume":"15","author":"Prisco","year":"2025","journal-title":"Diagnostics"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"104665","DOI":"10.1016\/j.dsp.2024.104665","article-title":"Gait recognition using deep learning with handling defective data from multiple wearable sensors","volume":"154","author":"Qin","year":"2024","journal-title":"Digit. Signal Process."},{"key":"ref39","doi-asserted-by":"publisher","first-page":"556","DOI":"10.3390\/a17120556","article-title":"Human activity recognition: a comparative study of validation methods and impact of feature extraction in wearable sensors","volume":"17","author":"Rehman","year":"2024","journal-title":"Algorithms"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1186\/s12984-021-00815-5","article-title":"Systematic review on wearable lower-limb exoskeletons for gait training in neuromuscular impairments","volume":"18","author":"Rodr\u00edguez-Fern\u00e1ndez","year":"2021","journal-title":"J. Neuroeng. Rehabil."},{"key":"ref41","author":"Scano","year":"2017"},{"key":"ref42","doi-asserted-by":"publisher","first-page":"3647","DOI":"10.3390\/s20133647","article-title":"Comparing person-specific and independent models on subject-dependent and independent human activity recognition performance","volume":"20","author":"Scheurer","year":"2020","journal-title":"Sensors"},{"key":"ref43","doi-asserted-by":"publisher","first-page":"3422","DOI":"10.1109\/JBHI.2024.3390606","article-title":"Understanding the role of self-attention in a transformer model for the discrimination of SCD from MCI using resting-state EEG","volume":"28","author":"Sibilano","year":"2024","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref44","doi-asserted-by":"publisher","first-page":"456","DOI":"10.1038\/s41551-022-00984-1","article-title":"Opportunities and challenges in the development of exoskeletons for locomotor assistance","volume":"7","author":"Siviy","year":"2023","journal-title":"Nature Biomed. Eng."},{"key":"ref45","doi-asserted-by":"publisher","first-page":"109653","DOI":"10.1016\/j.bspc.2026.109653","article-title":"Muscle networks analysis on an active occupational shoulder exoskeleton","volume":"119","author":"Suglia","year":"2026","journal-title":"Biomed. Signal Proc. Control."},{"key":"ref46","author":"Suglia","year":"2025"},{"key":"ref47","doi-asserted-by":"publisher","first-page":"2199","DOI":"10.3390\/s24072199","article-title":"A novel framework based on deep learning architecture for continuous human activity recognition with inertial sensors","volume":"24","author":"Suglia","year":"2024","journal-title":"Sensors"},{"key":"ref48","doi-asserted-by":"publisher","first-page":"108705","DOI":"10.1016\/j.engappai.2024.108705","article-title":"Efficient human activity recognition: a deep convolutional transformer-based contrastive self-supervised approach using wearable sensors","volume":"135","author":"Sun","year":"2024","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref49","doi-asserted-by":"publisher","first-page":"2390","DOI":"10.3390\/s24082390","article-title":"IMU-based real-time estimation of gait phase using multi-resolution neural networks","volume":"24","author":"Tang","year":"2024","journal-title":"Sensors"},{"key":"ref50","doi-asserted-by":"publisher","first-page":"13488","DOI":"10.1109\/JSEN.2024.3364187","article-title":"Large receptive field attention: an innovation in decomposing large-kernel convolution for sensor-based activity recognition","volume":"24","author":"Teng","year":"2024","journal-title":"IEEE Sensors J."},{"key":"ref51","doi-asserted-by":"publisher","first-page":"17746","DOI":"10.1109\/JIOT.2023.3277592","article-title":"Attention-based multihead deep learning framework for online activity monitoring with smartwatch sensors","volume":"10","author":"Thakur","year":"2023","journal-title":"IEEE Internet Things J."},{"key":"ref52","first-page":"5998","volume-title":"Attention Is All You Need: Advances in Neural Information Processing Systems","author":"Vaswani","year":"2017"},{"key":"ref53","doi-asserted-by":"publisher","first-page":"30","DOI":"10.3389\/fnbot.2017.00030","article-title":"An adaptive neuromuscular controller for assistive lower-limb exoskeletons: a preliminary study on subjects with spinal cord injury","volume":"11","author":"Wu","year":"2017","journal-title":"Front. Neurorobot."},{"key":"ref54","volume-title":"Proceedings of the 4th International Workshop on Sensor-based Activity Recognition and Interaction (iWOAR 2017)","author":"Yordanova","year":"2017"},{"key":"ref55","doi-asserted-by":"publisher","first-page":"115396","DOI":"10.1016\/j.measurement.2024.115396","article-title":"Human gait phases recognition based on multi-source data fusion and BiLSTM attention neural network","volume":"238","author":"Zhan","year":"2024","journal-title":"Measurement"},{"key":"ref56","doi-asserted-by":"publisher","first-page":"1476","DOI":"10.3390\/s22041417","article-title":"Deep learning in human activity recognition with wearable sensors: a review on advances","volume":"22","author":"Zhang","year":"2022","journal-title":"Sensors"},{"key":"ref57","doi-asserted-by":"publisher","first-page":"2377","DOI":"10.3390\/s24072377","article-title":"Recognition of human lower limb motion and muscle fatigue status based on wearable FES device and parallel structured deep learning","volume":"24","author":"Zhang","year":"2024","journal-title":"Sensors"},{"key":"ref58","doi-asserted-by":"publisher","first-page":"106417","DOI":"10.1016\/j.bspc.2024.106417","article-title":"Muscle synergies and muscle networks in multiple frequency components in post-stroke patients","volume":"95","author":"Zhao","year":"2024","journal-title":"Biomed. Signal Process. Control"}],"container-title":["Frontiers in Neurorobotics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2026.1901746\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T05:52:16Z","timestamp":1784699536000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2026.1901746\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,22]]},"references-count":54,"alternative-id":["10.3389\/fnbot.2026.1901746"],"URL":"https:\/\/doi.org\/10.3389\/fnbot.2026.1901746","relation":{},"ISSN":["1662-5218"],"issn-type":[{"value":"1662-5218","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,22]]},"article-number":"1901746"}}