Ling-Long Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ling-Long Li has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Guang-Zhong Cao
Shenzhen University
7 shared publications - Yue-Peng Zhang
Shenzhen University of Information Technology
4 shared publications Hong-Jie Liang
Guangdong Key Laboratory of Electromagnetic Control and Intelligent Robots, College of Mechatronics and Control Engineering, Shenzhen University, Shenzhen, China
2 shared publications- Wen-Zhou Li
Guangdong Key Laboratory of Electromagnetic Control and Intelligent Robots, College of Mechatronics and Control Engineering, Shenzhen University, Shenzhen 518060, China
2 shared publications Dongfeng Diao
1 shared publication- Dongfeng Diao
Institute of Nanosurface Science and Engineering Shenzhen University 518060 Shenzhen China
1 shared publication
Latest Publications
- 14 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (5) , 9056-9069
A Novel Lower Limb Motion Intention Recognition Model Using ConvNet With Prob-Sparse Attention
Motor imagery (MI)-electroencephalogram (EEG)-based brain-computer interfaces (BCIs) have emerged as transformative tools for enhancing clinical and real-life applications. Decoding MI-EEG signals to recognize lower-limb motion intentions (LLMIs) is highly important for controlling lower limb exoskeletons (LLEs) used in the rehabilitation of disabled individuals. Traditional approaches focus primarily on decoding upper-limb MI-EEG signals to indirectly infer LLMIs. However, recognizing direct LLMI to control LLEs is more acceptable and natural for users. This paper proposes a novel LLMI recognition model using convolutional neural networks with prob-sparse attention, to directly decode lower-limb MI-EEG signals for LLE control. This model synergizes a convolutional block for extracting localized temporal-spatial features with a prob-sparse attention block, which emphasizes distinctive global temporal features while reducing computational complexity. Offline experiments demonstrate that the proposed model achieves an average classification accuracy (ACA) exceeding 89.04%, outperforming existing state-of-the-art models. In online experiments, an EEG-based BCI system tailored for real-time LLE control is developed that features online data acquisition, model fine-tuning and EEG decoding capabilities, and the model achieves an ACA of 57.28% with an information transfer rate of 7.70 bits/min. The effectiveness and practicality of the proposed model for LLMI recognition are demonstrated in both offline and online experiments.
- 14 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (5) , 9056-9069
A Novel Lower Limb Motion Intention Recognition Model Using ConvNet With Prob-Sparse Attention
Motor imagery (MI)-electroencephalogram (EEG)-based brain-computer interfaces (BCIs) have emerged as transformative tools for enhancing clinical and real-life applications. Decoding MI-EEG signals to recognize lower-limb motion intentions (LLMIs) is highly important for controlling lower limb exoskeletons (LLEs) used in the rehabilitation of disabled individuals. Traditional approaches focus primarily on decoding upper-limb MI-EEG signals to indirectly infer LLMIs. However, recognizing direct LLMI to control LLEs is more acceptable and natural for users. This paper proposes a novel LLMI recognition model using convolutional neural networks with prob-sparse attention, to directly decode lower-limb MI-EEG signals for LLE control. This model synergizes a convolutional block for extracting localized temporal-spatial features with a prob-sparse attention block, which emphasizes distinctive global temporal features while reducing computational complexity. Offline experiments demonstrate that the proposed model achieves an average classification accuracy (ACA) exceeding 89.04%, outperforming existing state-of-the-art models. In online experiments, an EEG-based BCI system tailored for real-time LLE control is developed that features online data acquisition, model fine-tuning and EEG decoding capabilities, and the model achieves an ACA of 57.28% with an information transfer rate of 7.70 bits/min. The effectiveness and practicality of the proposed model for LLMI recognition are demonstrated in both offline and online experiments.
- Open Access
MACNet: A Multidimensional Attention-Based Convolutional Neural Network for Lower-Limb Motor Imagery Classification
Decoding lower-limb motor imagery (MI) is highly important in brain–computer interfaces (BCIs) and rehabilitation engineering. However, it is challenging to classify lower-limb MI from electroencephalogram (EEG) signals, because lower-limb motions (LLMs) including MI are excessively close to physiological representations in the human brain and generate low-quality EEG signals. To address this challenge, this paper proposes a multidimensional attention-based convolutional neural network (CNN), termed MACNet, which is specifically designed for lower-limb MI classification. MACNet integrates a temporal refining module and an attention-enhanced convolutional module by leveraging the local and global feature representation abilities of CNNs and attention mechanisms. The temporal refining module adaptively investigates critical information from each electrode channel to refine EEG signals along the temporal dimension. The attention-enhanced convolutional module extracts temporal and spatial features while refining the feature maps across the channel and spatial dimensions. Owing to the scarcity of public datasets available for lower-limb MI, a specified lower-limb MI dataset involving four routine LLMs is built, consisting of 10 subjects over 20 sessions. Comparison experiments and ablation studies are conducted on this dataset and a public BCI Competition IV 2a EEG dataset. The experimental results show that MACNet achieves state-of-the-art performance and outperforms alternative models for the subject-specific mode. Visualization analysis reveals the excellent feature learning capabilities of MACNet and the potential relationship between lower-limb MI and brain activity. The effectiveness and generalizability of MACNet are verified.
Open AccessMACNet: A Multidimensional Attention-Based Convolutional Neural Network for Lower-Limb Motor Imagery Classification
Decoding lower-limb motor imagery (MI) is highly important in brain–computer interfaces (BCIs) and rehabilitation engineering. However, it is challenging to classify lower-limb MI from electroencephalogram (EEG) signals, because lower-limb motions (LLMs) including MI are excessively close to physiological representations in the human brain and generate low-quality EEG signals. To address this challenge, this paper proposes a multidimensional attention-based convolutional neural network (CNN), termed MACNet, which is specifically designed for lower-limb MI classification. MACNet integrates a temporal refining module and an attention-enhanced convolutional module by leveraging the local and global feature representation abilities of CNNs and attention mechanisms. The temporal refining module adaptively investigates critical information from each electrode channel to refine EEG signals along the temporal dimension. The attention-enhanced convolutional module extracts temporal and spatial features while refining the feature maps across the channel and spatial dimensions. Owing to the scarcity of public datasets available for lower-limb MI, a specified lower-limb MI dataset involving four routine LLMs is built, consisting of 10 subjects over 20 sessions. Comparison experiments and ablation studies are conducted on this dataset and a public BCI Competition IV 2a EEG dataset. The experimental results show that MACNet achieves state-of-the-art performance and outperforms alternative models for the subject-specific mode. Visualization analysis reveals the excellent feature learning capabilities of MACNet and the potential relationship between lower-limb MI and brain activity. The effectiveness and generalizability of MACNet are verified.
- Open Access
- 21 November 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (11) , e0309706
FDCN-C: A deep learning model based on frequency enhancement, deformable convolution network, and crop module for electroencephalography motor imagery classification
Motor imagery (MI)-electroencephalography (EEG) decoding plays an important role in brain-computer interface (BCI), which enables motor-disabled patients to communicate with external world via manipulating smart equipment. Currently, deep learning (DL)-based methods are popular for EEG decoding. Whereas the utilization efficiency of EEG features in frequency and temporal domain is not sufficient, which results in poor MI classification performance. To address this issue, an EEG-based MI classification model based on a frequency enhancement module, a deformable convolutional network, and a crop module (FDCN-C) is proposed. Firstly, the frequency enhancement module is innovatively designed to address the issue of extracting frequency information. It utilizes convolution kernels at continuous time scales to extract features across different frequency bands. These features are screened by calculating attention and integrated into the original EEG data. Secondly, for temporal feature extraction, a deformable convolution network is employed to enhance feature extraction capabilities, utilizing offset parameters to modulate the convolution kernel size. In spatial domain, a one-dimensional convolution layer is designed to integrate all channel information. Finally, a dilated convolution is used to form a crop classification module, wherein the diverse receptive fields of the EEG data are computed multiple times. Two public datasets are employed to verify the proposed FDCN-C model, the classification accuracy obtained from the proposed model is greater than that of state-of-the-art methods. The model’s accuracy has improved by 14.01% compared to the baseline model, and the ablation study has confirmed the effectiveness of each module in the model.
- 21 November 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (11) , e0309706
Open AccessFDCN-C: A deep learning model based on frequency enhancement, deformable convolution network, and crop module for electroencephalography motor imagery classification
Motor imagery (MI)-electroencephalography (EEG) decoding plays an important role in brain-computer interface (BCI), which enables motor-disabled patients to communicate with external world via manipulating smart equipment. Currently, deep learning (DL)-based methods are popular for EEG decoding. Whereas the utilization efficiency of EEG features in frequency and temporal domain is not sufficient, which results in poor MI classification performance. To address this issue, an EEG-based MI classification model based on a frequency enhancement module, a deformable convolutional network, and a crop module (FDCN-C) is proposed. Firstly, the frequency enhancement module is innovatively designed to address the issue of extracting frequency information. It utilizes convolution kernels at continuous time scales to extract features across different frequency bands. These features are screened by calculating attention and integrated into the original EEG data. Secondly, for temporal feature extraction, a deformable convolution network is employed to enhance feature extraction capabilities, utilizing offset parameters to modulate the convolution kernel size. In spatial domain, a one-dimensional convolution layer is designed to integrate all channel information. Finally, a dilated convolution is used to form a crop classification module, wherein the diverse receptive fields of the EEG data are computed multiple times. Two public datasets are employed to verify the proposed FDCN-C model, the classification accuracy obtained from the proposed model is greater than that of state-of-the-art methods. The model’s accuracy has improved by 14.01% compared to the baseline model, and the ablation study has confirmed the effectiveness of each module in the model.
- Open Access
Human-in-the-Loop Trajectory Optimization Based on sEMG Biofeedback for Lower-Limb Exoskeleton
Lower-limb exoskeletons (LLEs) can provide rehabilitation training and walking assistance for individuals with lower-limb dysfunction or those in need of functionality enhancement. Adapting and personalizing the LLEs is crucial for them to form an intelligent human–machine system (HMS). However, numerous LLEs lack thorough consideration of individual differences in motion planning, leading to subpar human performance. Prioritizing human physiological response is a critical objective of trajectory optimization for the HMS. This paper proposes a human-in-the-loop (HITL) motion planning method that utilizes surface electromyography signals as biofeedback for the HITL optimization. The proposed method combines offline trajectory optimization with HITL trajectory selection. Based on the derived hybrid dynamical model of the HMS, the offline trajectory is optimized using a direct collocation method, while HITL trajectory selection is based on Thompson sampling. The direct collocation method optimizes various gait trajectories and constructs a gait library according to the energy optimality law, taking into consideration dynamics and walking constraints. Subsequently, an optimal gait trajectory is selected for the wearer using Thompson sampling. The selected gait trajectory is then implemented on the LLE under a hybrid zero dynamics control strategy. Through the HITL optimization and control experiments, the effectiveness and superiority of the proposed method are verified.
Open AccessHuman-in-the-Loop Trajectory Optimization Based on sEMG Biofeedback for Lower-Limb Exoskeleton
Lower-limb exoskeletons (LLEs) can provide rehabilitation training and walking assistance for individuals with lower-limb dysfunction or those in need of functionality enhancement. Adapting and personalizing the LLEs is crucial for them to form an intelligent human–machine system (HMS). However, numerous LLEs lack thorough consideration of individual differences in motion planning, leading to subpar human performance. Prioritizing human physiological response is a critical objective of trajectory optimization for the HMS. This paper proposes a human-in-the-loop (HITL) motion planning method that utilizes surface electromyography signals as biofeedback for the HITL optimization. The proposed method combines offline trajectory optimization with HITL trajectory selection. Based on the derived hybrid dynamical model of the HMS, the offline trajectory is optimized using a direct collocation method, while HITL trajectory selection is based on Thompson sampling. The direct collocation method optimizes various gait trajectories and constructs a gait library according to the energy optimality law, taking into consideration dynamics and walking constraints. Subsequently, an optimal gait trajectory is selected for the wearer using Thompson sampling. The selected gait trajectory is then implemented on the LLE under a hybrid zero dynamics control strategy. Through the HITL optimization and control experiments, the effectiveness and superiority of the proposed method are verified.
- 6 March 2024
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 93
A Fast Calibration Method for an sEMG-Based Lower Limb Joint Torque Estimation Model
- 6 March 2024
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 93
A Fast Calibration Method for an sEMG-Based Lower Limb Joint Torque Estimation Model
- 10 August 2022
- Published by Springer Nature
- p. 386-399
EEG Generation of Virtual Channels Using an Improved Wasserstein Generative Adversarial Networks
- 10 August 2022
- Published by Springer Nature
- p. 386-399
EEG Generation of Virtual Channels Using an Improved Wasserstein Generative Adversarial Networks
- Open Access
- 3 November 2021
- Published by MDPI AG in Applied Sciences
- Vol. 11 (21) , 10329
A Self-Adaptive-Coefficient-Double-Power Sliding Mode Control Method for Lower Limb Rehabilitation Exoskeleton Robot
Lower limb rehabilitation exoskeleton robots have the characteristics of nonlinearity and strong coupling, and they are easily disturbed during operation by environmental factors. Thus, an accurate dynamic model of the robot is difficult to obtain, and achieving trajectory tracking control of the robot is also difficult. In this article, a self-adaptive-coefficient double-power sliding mode control method is proposed to overcome the difficulty of tracking the robot trajectory. The method combines an estimated dynamic model with sliding mode control. A nonlinear control law was designed based on the robot dynamics model and computational torque method, and a compensation term of control law based on double-power reaching law was introduced to reduce the disturbance from model error and environmental factors. The self-adaptive coefficient of the compensation term of the control law was designed to adaptively adjust the compensation term to improve the anti-interference ability of the robot. The simulation and experiment results show that the proposed method effectively improves the trajectory tracking accuracy and anti-interference ability of the robot. Compared with the traditional computed torque method, the proposed method decreases the tracking error by more than 71.77%. The maximum absolute error of the hip joint and knee joint remained below 0.55° and 1.65°, respectively, in the wearable experiment of the robot.
- 3 November 2021
- Published by MDPI AG in Applied Sciences
- Vol. 11 (21) , 10329
Open AccessA Self-Adaptive-Coefficient-Double-Power Sliding Mode Control Method for Lower Limb Rehabilitation Exoskeleton Robot
Lower limb rehabilitation exoskeleton robots have the characteristics of nonlinearity and strong coupling, and they are easily disturbed during operation by environmental factors. Thus, an accurate dynamic model of the robot is difficult to obtain, and achieving trajectory tracking control of the robot is also difficult. In this article, a self-adaptive-coefficient double-power sliding mode control method is proposed to overcome the difficulty of tracking the robot trajectory. The method combines an estimated dynamic model with sliding mode control. A nonlinear control law was designed based on the robot dynamics model and computational torque method, and a compensation term of control law based on double-power reaching law was introduced to reduce the disturbance from model error and environmental factors. The self-adaptive coefficient of the compensation term of the control law was designed to adaptively adjust the compensation term to improve the anti-interference ability of the robot. The simulation and experiment results show that the proposed method effectively improves the trajectory tracking accuracy and anti-interference ability of the robot. Compared with the traditional computed torque method, the proposed method decreases the tracking error by more than 71.77%. The maximum absolute error of the hip joint and knee joint remained below 0.55° and 1.65°, respectively, in the wearable experiment of the robot.