{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T15:18:26Z","timestamp":1783005506430,"version":"3.54.5"},"reference-count":29,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,6,25]],"date-time":"2026-06-25T00:00:00Z","timestamp":1782345600000},"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. Comput. Neurosci."],"abstract":"<jats:p>To enhance the dynamic adaptability and robustness of neuromuscular signal-driven exoskeleton robots in complex environments. This study develops a Deep Reinforcement Learning (DRL)-based exoskeleton control framework. A module for neuromuscular signal processing and motion intention modeling is designed, encompassing band-pass filtering, local normalization, time-frequency feature extraction, and Bidirectional Long Short-Term Memory (BiLSTM) time-series encoding. Additionally, a DRL-based dynamically adaptive control framework is established. Trajectory tracking error, torque smoothness, energy consumption agent, joint safety constraint and intention consistency are comprehensively introduced into the reward function to achieve the multi-objective balance of \u201caccuracy-comfort-energy consumption-safety.\u201d Comparative experiments demonstrate that the joint trajectory tracking errors of the proposed optimized system are 2.37\u00b0, 2.64\u00b0, and 2.92\u00b0, respectively. These values are significantly lower than those of comparative systems, including the Deep Reinforcement Learning-based Robust Controller for Lower-Limb Rehabilitation Exoskeletons (DRL-RC-LLRE) and the Learning-in-Simulation Exoskeleton Assistance Framework (LiS-EXO). The corresponding torque smoothness indicators are 0.62, 0.68, and 0.71. The mechanical energy consumption proxy indicators are 10.48, 11.39, and 11.67, indicating that while improving the tracking accuracy, the torque oscillation is effectively suppressed and the energy consumption is reduced. The ablation experiments further show that when any of the neuromuscular intention embedding, intention consistency reward or safety constraint modules are removed, the average trajectory error, torque smoothness and energy consumption indicators all deteriorate to varying degrees. The complete model achieves 2.63\u00b0, 0.66, and 11.17, respectively, in the three indicators, verifying the collaborative contribution of the three key modules to the overall performance. Overall, this study has constructed and verified a reproducible \u201cneuromuscular signal-motion intention-reinforcement learning control\u201d integrated framework, and thus having certain contributions to the field of intelligent exoskeleton control.<\/jats:p>","DOI":"10.3389\/fncom.2026.1807958","type":"journal-article","created":{"date-parts":[[2026,6,25]],"date-time":"2026-06-25T04:22:13Z","timestamp":1782361333000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Design of dynamic adaptive control framework for exoskeleton robot driven by neuromuscular signal based on deep reinforcement learning"],"prefix":"10.3389","volume":"20","author":[{"given":"Kaidi","family":"Ma","sequence":"first","affiliation":[{"name":"AI Institute, Jiangsu XCMG National Key Laboratory Technology Co., Ltd.","place":["Xuzhou, Jiangsu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,6,25]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"40895","DOI":"10.1038\/s41598-025-24831-w","article-title":"A hybrid EMG\u2013EEG interface for robust intention detection and fatigue-adaptive control of an elbow rehabilitation robot","volume":"15","author":"Abdallah","year":"2025","journal-title":"Sci. Rep."},{"key":"ref2","doi-asserted-by":"publisher","first-page":"331","DOI":"10.3390\/bioengineering12040331","article-title":"Deep learning-enhanced motor training: a hybrid VR and exoskeleton system for cognitive\u2013motor rehabilitation","volume":"12","author":"Acu\u00f1a Luna","year":"2025","journal-title":"Bioengineering"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"e0320211","DOI":"10.1371\/journal.pone.0320211","article-title":"Characterization of human balance through a reinforcement learning-based muscle controller","volume":"20","author":"Akba\u015f","year":"2025","journal-title":"PLoS One"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"665","DOI":"10.3390\/biomimetics9110665","article-title":"Comparison of empirical and reinforcement learning (RL)-based control based on proximal policy optimization (PPO) for walking assistance: does AI always win?","volume":"9","author":"Drewing","year":"2024","journal-title":"Biomimetics"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"397","DOI":"10.1007\/s11044-023-09884-x","article-title":"Multibody dynamics and control using machine learning","volume":"58","author":"Hashemi","year":"2023","journal-title":"Multibody Syst. Dyn."},{"key":"ref6","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1007\/s10846-025-02230-7","article-title":"A lower limb exoskeleton adaptive control method based on model-free reinforcement learning and improved dynamic movement primitives","volume":"111","author":"Huang","year":"2025","journal-title":"J. Intell. Robot. Syst."},{"key":"ref7","doi-asserted-by":"publisher","first-page":"460","DOI":"10.3390\/biomimetics10070460","article-title":"AI-driven control strategies for biomimetic robotics: trends, challenges, and future directions","volume":"10","author":"Jung","year":"2025","journal-title":"Biomimetics"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"10054","DOI":"10.1109\/jsen.2023.3259150","article-title":"Implementing hand gesture recognition using EMG on the Zynq circuit","volume":"23","author":"Kerdjidj","year":"2023","journal-title":"IEEE Sensors J."},{"key":"ref9","first-page":"7","article-title":"AI-enhanced robotic exoskeletons using reinforcement learning for personalized motion assistance","volume":"1","author":"Khan","year":"2025","journal-title":"Int. J. Adv. Innov. Res."},{"key":"ref10","doi-asserted-by":"publisher","first-page":"3675","DOI":"10.1109\/tsmc.2024.3369071","article-title":"A survey of wearable lower extremity neurorehabilitation exoskeleton: sensing, gait dynamics, and human\u2013robot collaboration","volume":"54","author":"Li","year":"2024","journal-title":"IEEE Trans Syst Man Cybern Syst"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1186\/s12984-023-01147-2","article-title":"Robust walking control of a lower limb rehabilitation exoskeleton coupled with a musculoskeletal model via deep reinforcement learning","volume":"20","author":"Luo","year":"2023","journal-title":"J. Neuroeng. Rehabil."},{"key":"ref12","doi-asserted-by":"publisher","first-page":"11","DOI":"10.64206\/jbrwaf31","article-title":"Gait analysis and rehabilitation robotics: a hybrid approach using wearable sensors and reinforcement learning","volume":"2","author":"M\u00fcller","year":"2024","journal-title":"Stem Cell Artif. Intell. Data Sci. J."},{"key":"ref13","doi-asserted-by":"publisher","first-page":"270","DOI":"10.1007\/s42235-023-00453-8","article-title":"Robust machine learning mapping of sEMG signals to future actuator commands in biomechatronic devices","volume":"21","author":"Nasr","year":"2024","journal-title":"J. Bionic Eng."},{"key":"ref14","doi-asserted-by":"publisher","first-page":"9184","DOI":"10.3390\/app15169184","article-title":"Control algorithms in robot-assisted rehabilitation: a systematic review","volume":"15","author":"Rad","year":"2025","journal-title":"Appl. Sci."},{"key":"ref15","doi-asserted-by":"publisher","first-page":"106","DOI":"10.3390\/biomimetics10020106","article-title":"Long short-term memory-enabled electromyography-controlled adaptive wearable robotic exoskeleton for upper arm rehabilitation","volume":"10","author":"Samarakoon","year":"2025","journal-title":"Biomimetics"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1109\/LRA.2023.3330678","article-title":"Emg-based intention detection using deep learning for shared control in upper-limb assistive exoskeletons","volume":"9","author":"Sedighi","year":"2023","journal-title":"IEEE Robot. Autom. Lett."},{"key":"ref17","doi-asserted-by":"publisher","first-page":"3528","DOI":"10.3390\/s25113528","article-title":"Enhancing upper limb exoskeletons using sensor-based deep learning torque prediction and PID control","volume":"25","author":"Shakeriaski","year":"2025","journal-title":"Sensors"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"5778","DOI":"10.3390\/s23135778","article-title":"BioMAT: an open-source biomechanics multi-activity transformer for joint kinematic predictions using wearable sensors","volume":"23","author":"Sharifi-Renani","year":"2023","journal-title":"Sensors"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"9546","DOI":"10.3390\/app13179546","article-title":"Review of sEMG for robot control: techniques and applications","volume":"13","author":"Song","year":"2023","journal-title":"Appl. Sci."},{"key":"ref20","doi-asserted-by":"publisher","first-page":"4004","DOI":"10.3390\/s25134004","article-title":"Electromyography signal acquisition, filtering, and data analysis for exoskeleton development","volume":"25","author":"Sul","year":"2025","journal-title":"Sensors"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"214","DOI":"10.3390\/biomimetics10040214","article-title":"Myoelectric control in rehabilitative and assistive soft exoskeletons: a comprehensive review of trends, challenges, and integration with soft robotic devices","volume":"10","author":"Toro-Ossaba","year":"2025","journal-title":"Biomimetics"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"7329","DOI":"10.1126\/scirobotics.adt7329","article-title":"AI in therapeutic and assistive exoskeletons and exosuits: influences on performance and autonomy","volume":"10","author":"Van der Kooij","year":"2025","journal-title":"Sci. Robot."},{"key":"ref23","doi-asserted-by":"publisher","first-page":"465","DOI":"10.20517\/ais.2025.03","article-title":"Machine learning for regenerative peripheral nerve interface-based prosthetic control: current applications and clinical translation","volume":"5","author":"Wang","year":"2025","journal-title":"Artif. Intell. Surg."},{"key":"ref24","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s10489-025-06360-1","article-title":"AI-computing, deep reinforcement learning-based predictive human-robot neuromechanical simulation for wearable robots","volume":"55","author":"Wang","year":"2025","journal-title":"Appl. Intell."},{"key":"ref25","doi-asserted-by":"publisher","first-page":"1075","DOI":"10.1109\/jas.2024.124329","article-title":"Intuitive human-robot-environment interaction with EMG signals: a review","volume":"11","author":"Xiong","year":"2024","journal-title":"IEEE\/CAA J. Autom. Sin."},{"key":"ref26","first-page":"32","article-title":"Deep reinforcement learning in healthcare robotics: a pathway towards intelligent assistive care","volume":"2","author":"Yousuf","year":"2025","journal-title":"Front. Multidiscip. Stud."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"82","DOI":"10.1007\/s11424-024-3414-7","article-title":"A bio-inspired integration model of basal ganglia and cerebellum for motion learning of a musculoskeletal robot","volume":"37","author":"Zhang","year":"2024","journal-title":"J. Syst. Sci. Complex."},{"key":"ref28","doi-asserted-by":"publisher","first-page":"2448","DOI":"10.3390\/s25082448","article-title":"Review of sEMG for exoskeleton robots: motion intention recognition techniques and applications","volume":"25","author":"Zhang","year":"2025","journal-title":"Sensors"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"194","DOI":"10.1186\/s10033-025-01355-y","article-title":"Virtual impedance adaptation of lower-limb exoskeleton for human performance augmentation based on deep reinforcement learning","volume":"38","author":"Zheng","year":"2025","journal-title":"Chin. J. Mech. Eng."}],"container-title":["Frontiers in Computational Neuroscience"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2026.1807958\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T14:57:58Z","timestamp":1783004278000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fncom.2026.1807958\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,25]]},"references-count":29,"alternative-id":["10.3389\/fncom.2026.1807958"],"URL":"https:\/\/doi.org\/10.3389\/fncom.2026.1807958","relation":{},"ISSN":["1662-5188"],"issn-type":[{"value":"1662-5188","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,25]]},"article-number":"1807958"}}