Lijun Han

Lijun Han

Auto-generated profile, based on publicly available information

Lijun Han

School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
9 September 2024 in IEEE Transactions on Biomedical Engineering
Verified
Journal Article
Closed access
LHLijun Han
Lijun Han
LCLong Cheng
Long Cheng
HLHoucheng Li
Houcheng Li
YZYongxiang Zou
Yongxiang Zou
SQShijie Qin
Shijie Qin
MZMing Zhou
Ming Zhou
Objective : Surface electromyography (sEMG) driven musculoskeletal models are promising to be applied in the field of human-computer interaction. However, due to the individual-specific physiological characteristics, generic models often fail to provide accurate motion estimation. This study optimized the general model to build a personalized model and improve the accuracy of motion estimation. Methods : Inspired by the coupling effect of wrist/hand movement, a hierarchical optimization approach for personalizing musculoskeletal models (HOPE-MM) is proposed, which aligns with the physiological characteristics of the human wrist and hand. To verify the effectiveness of personalized musculoskeletal model, single joint motions and simultaneous joint motions are estimated. In addition, Sobol sensitivity analysis is conducted to identify the key parameters of musculoskeletal model, providing guidance for model simplification. Results : The mean pearson correlation coefficient between the predicted joint angles and the measured joint angles are 0.95 $\pm$ 0.03 and 0.93 $\pm$ 0.01 for simultaneous wrist and metacarpophalangeal (MCP) joint movements, respectively, which have a significant improvement compared with the stateof- the-art works. By optimizing only the key parameters including tendon slack length, maximal isometric force and optimal fiber length, the performances of simplified model are comparable to the full-parameter model. Conclusion : These results provide insights into the effects of muscletendon parameters on musculoskeletal model, and musculoskeletal models personalized using hierarchical optimization methods can improve the accuracy of motion estimates. Significance: These findings facilitate the clinical application of musculoskeletal models in rehabilitation and robotic control.

21 May 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
LHLijun Han
Lijun Han
LCLong Cheng
Long Cheng
YZYongxiang Zou
Yongxiang Zou
YLYanan Li
Yanan Li
Accurately estimating joint angles and torques is essential for human-machine interfaces, such as exoskeleton control and human-robot collaboration. Musculoskeletal models driven by surface electromyographic (sEMG) signals incorporate human physiological knowledge but are computationally complex, whereas data-driven methods offer faster inference but lack interpretability. To bridge this gap, a physics-informed deep transfer learning framework is proposed in this article, which integrates physical knowledge as a soft constraint into the data-driven method. Additionally, a novel neural network called convolutional neural network (CNN)–Mamba is proposed to implement the framework. It combines the local feature extraction capability of the CNN with the global modeling ability of the state-space model (SSM) known as Mamba. To demonstrate its capability to simultaneously estimate the angles and torques for multiple joints, the method was validated on two datasets: wrist-hand coupled motion (ten subjects, a total of 30 trials) and walking motion (eight subjects, a total of 32 trials). The average Pearson correlation coefficient between estimated values and ground truth for all joints exceeded 0.96, reaching a maximum of 0.99. Ablation and comparative experiments further confirmed the advantages of the proposed method. Moreover, our approach demonstrates comparable performance to model-based methods while achieving at least $14\times $ faster inference speed. Specifically, estimating the joint angle and torque for a motion duration of 1 s takes only 161.8 ms. Overall, our innovative approach provides an effective solution for estimating multiphysiological parameters across multiple joints in human motion, with promising applications in human-machine interaction (HMI).

27 November 2024 in IEEE/ASME Transactions on Mechatronics
Verified
Journal Article
Closed access
XXXiuze Xia
Xiuze Xia
LCLong Cheng
Long Cheng
MMMuyuan Ma
Muyuan Ma
HZHaoyu Zhang
Haoyu Zhang
LHLijun Han
Lijun Han
HLHoucheng Li
Houcheng Li
One of the urgent issues with wearable exoskeletons is their difficulty in following human fast movements, especially nonperiodic arm movements. Wearing an exoskeleton can make human movement stiff, which impacts its long-term usability. To mitigate this effect, this article proposes a novel neurologically inspired transparent interaction strategy for arm exoskeletons, which employs a hierarchical controller. First, inspired by the neurological concept of vigor, we propose a top-level predictor to predict arm fast movements in real time. In addition, we incorporate a bidirectional human–robot adaptation strategy into the predictor to enhance the integration of human and exoskeleton. Furthermore, a novel model predictive variable impedance controller is proposed as the mid-level and low-level controllers. We incorporate the interaction force into the weight matrix of the cost function used in model predictive control to further reduce the human–robot interaction force. Experiments using metrics, such as interaction force, surface electromyography, and changes in movement habits, demonstrated that the proposed transparent interaction strategy can effectively reduce the impact of the exoskeleton on users.

6 June 2025 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
HLHoucheng Li
Houcheng Li
LCLong Cheng
Long Cheng
SQShijie Qin
Shijie Qin
LHLijun Han
Lijun Han
This paper presents a novel voluntary control method for a hand assistive exoskeleton, leveraging an sEMG-driven musculoskeletal model to improve the performance of grasping tasks. To address the challenge of inadequate personalization in current hand exoskeleton assistance strategies, this research first establishes a musculoskeletal model based on the Hill-type framework. Through experimental calibration, a personalized musculoskeletal model is then developed, tailored to individual subjects. Subsequently, a method for estimating finger joint torque is introduced, leveraging the sEMG-driven personalized musculoskeletal model to achieve precise control. An exoskeleton controller based on admittance control is also designed, enabling users to voluntarily control the exoskeleton. By assessing the subject's voluntary effort, the proposed control method dynamically adjusts the level of exoskeleton assistance during grasping tasks. Finally, the effectiveness of both the personalized model and the voluntary control method is validated through rigorous experimental testing, demonstrating significant improvements in assistance strategy personalization and user experience.

1 October 2026 in IEEE Transactions on Biomedical Engineering
Verified
Journal Article
Closed access
LHLijun Han
Lijun Han
LCLong Cheng
Long Cheng
MMMuyuan Ma
Muyuan Ma
FYFeiyang Yang
Feiyang Yang
HLHoucheng Li
Houcheng Li
Background: Acquiring human- like motor skills in embodied musculoskeletal models is challenging due to the high dimensionality and redundancy of muscle actuators. Methods: Inspired by human motor control coordination patterns, we introduce a synergy-guided reinforcement learning framework integrating physiological priors derived from muscle synergies into the control policy of an embodied musculoskeletal model. It leverages coordinated muscle activation patterns to guide learning, generating muscle excitation signals via a synergy-guided control component and a residual control component. To evaluate the proposed method, four badminton stroke skills are selected as benchmark tasks (forehand/backhand, inward/outward net slices). Results: The experimental results demonstrate that our method achieves an average root mean square error of under 0.015 radians across all stroke types, demonstrating its ability to accurately learn expert motion. Furthermore, it outperforms the baseline proximal policy optimization (PPO) model in terms of trajectory accuracy, energy efficiency, and training convergence speed, particularly excelling in energy efficiency with up to a 14.9% reduction in energy consumption. The forehand high serve is also tested to validate the method's effectiveness in learning longer and larger-ranges movements, showing the same advantages. Moreover, the muscle synergies learned by the model exhibit moderate resemblance to human synergies, indicating potential interpretability and biological plausibility. Conclusion: This work highlights that integrating neurophysiological priors into reinforcement learning provides a promising pathway for efficient, interpretable, and human- like motor control. Significance: The approach holds promise for advancing motor skill assessment, human–machine interfaces, and rehabilitation technologies by enabling more efficient and human- like motion skill learning.

9 March 2026 in IEEE Transactions on Automation Science and Engineering
Verified
Journal Article
Closed access
SLSiyuan Liu
Siyuan Liu
BJBo Jiang
Bo Jiang
LHLijun Han
Lijun Han
SZShanlin Zhong
Shanlin Zhong
CSCi Song
Ci Song
JCJiahao Chen
Jiahao Chen
Musculoskeletal robots with anthropomorphic structures show great potential for human-like dexterous manipulation. However, their high-dimensional, coupled and nonlinear dynamics pose significant challenges to control and manipulation. Despite recent advances, most existing methods still struggle to achieve flexible multi-object manipulation without long-sequence expert demonstrations or manual parameter tuning based on expert knowledge. To tackle these limitations, a bio-inspired reinforcement learning framework named BioSynGrasp is proposed. It comprises three core components: 1) A neuromuscular controller that integrates shared perception and planning modules with distinct arm and hand synergies sub-networks; 2) A proximal policy optimization-based exploration strategy driven by hand and arm muscle synergies with correlated actuator perturbations; 3) Hand preshapes measured via motion capture provide efficient initialization for functional grasping and reaching with minimal data. BioSynGrasp is validated on a simulated musculoskeletal robot with 63 actuated muscles and on a hardware articulated dexterous hand system. Results demonstrate that it successfully completes 28 multi-object grasping and reaching tasks with an average success rate of 85.9%, outperforming baseline methods in success rate, energy efficiency, position error, and redundant actuator recruitment. These results highlight that integrating bio-inspired motor control with hand preshapes enhances human-like dexterous manipulation and provides insights into human neuromechanics for more adaptable robotic systems.