A Robot-Driven Computational Model for Estimating Passive Ankle Torque With Subject-Specific Adaptation
TL;DR: A computational ankle model is proposed for use in robot-assisted therapy with three rotational degrees of freedom, 12 muscles, and seven ligaments that uses three independent position variables as inputs, and outputs an overall ankle assessment.
Abstract: Background : Robot-assisted ankle assessment could potentially be conducted using sensor-based and model-based methods. Existing ankle rehabilitation robots usually use torquemeters and multiaxis load cells for measuring joint dynamics. These measurements are accurate, but the contribution as a result of muscles and ligaments is not taken into account. Some computational ankle models have been developed to evaluate ligament strain and joint torque. These models do not include muscles and, thus, are not suitable for an overall ankle assessment in robot-assisted therapy. Methods : This study proposed a computational ankle model for use in robot-assisted therapy with three rotational degrees of freedom, 12 muscles, and seven ligaments. This model is driven by robotics, uses three independent position variables as inputs, and outputs an overall ankle assessment. Subject-specific adaptations by geometric and strength scaling were also made to allow for a universal model. Results : This model was evaluated using published results and experimental data from 11 participants. Results show a high accuracy in the evaluation of ligament neutral length and passive joint torque. The subject-specific adaptation performance is high, with each normalized root-mean-square deviation value less than 10%. Conclusion : This model could be used for ankle assessment, especially in evaluating passive ankle torque, for a specific individual. The characteristic that is unique to this model is the use of three independent position variables that can be measured in real time as inputs, which makes it advantageous over other models when combined with robot-assisted therapy.
Citations
More filters
TL;DR: In this paper, a robust iterative feedback tuning (IFT) technique for repetitive training control of a compliant parallel ankle rehabilitation robot is presented, which employs four compliant pneumatic muscle actuators that mimic skeletal muscles for ankle's motion training.
Abstract: Robot-assisted rehabilitation offers benefits, such as repetitive, intensive, and task-specific training, as compared to traditional manual manipulation performed by physiotherapists. In this paper, a robust iterative feedback tuning (IFT) technique for repetitive training control of a compliant parallel ankle rehabilitation robot is presented. The robot employs four parallel intrinsically compliant pneumatic muscle actuators that mimic skeletal muscles for ankle's motion training. A multiple degrees-of-freedom normalized IFT technique is proposed to increase the controller robustness by obtaining an optimal value for the weighting factor and offering a method with learning capacity to achieve an optimum of the controller parameters. Experiments with human participants were conducted to investigate the robustness as well as to validate the performance of the proposed IFT technique. Results show that the normalized IFT scheme will achieve a better and better tracking performance during the robot repetitive control and provides more robustness to the system by adapting to various situations in robotic rehabilitation.
93 citations
Cites methods from "A Robot-Driven Computational Model ..."
...The ankle movement range and active torque of each participant were assessed first by using the method presented in our previous work [35], [36], to ensure that...
[...]
TL;DR: Experimental findings suggest the potential of this new adaptive patient-cooperative control strategy as a safe and engaging control solution for rehabilitation robots.
Abstract: This paper proposes a new adaptive patient-cooperative control strategy for improving the effectiveness and safety of robot-assisted ankle rehabilitation. This control strategy has been developed and implemented on a compliant ankle rehabilitation robot (CARR). The CARR is actuated by four Festo Fluidic muscles located to the calf in parallel, has three rotational degrees of freedom. The control scheme consists of a position controller implemented in joint space and a high-level admittance controller in task space. The admittance controller adaptively modifies the predefined trajectory based on real-time ankle measurement, which enhances the training safety of the robot. Experiments were carried out using different modes to validate the proposed control strategy on the CARR. Three training modes include: 1) a passive mode using a joint-space position controller, 2) a patient–robot cooperative mode using a fixed-parameter admittance controller, and 3) a cooperative mode using a variable-parameter admittance controller. Results demonstrate satisfactory trajectory tracking accuracy, even when externally disturbed, with a maximum normalized root mean square deviation less than 5.4%. These experimental findings suggest the potential of this new patient-cooperative control strategy as a safe and engaging control solution for rehabilitation robots.
80 citations
TL;DR: This article proposes a multichannel-based generative adversarial network (MGAN) with semisupervision to grade DR and demonstrates that the proposed model outperforms the other representative models in terms of accuracy, area under ROC curve (AUC), sensitivity, and specificity.
Abstract: Diabetic retinopathy (DR) is one of the major causes of blindness. It is of great significance to apply deep-learning techniques for DR recognition. However, deep-learning algorithms often depend on large amounts of labeled data, which is expensive and time-consuming to obtain in the medical imaging area. In addition, the DR features are inconspicuous and spread out over high-resolution fundus images. Therefore, it is a big challenge to learn the distribution of such DR features. This article proposes a multichannel-based generative adversarial network (MGAN) with semisupervision to grade DR. The multichannel generative model is developed to generate a series of subfundus images corresponding to the scattering DR features. By minimizing the dependence on labeled data, the proposed semisupervised MGAN can identify the inconspicuous lesion features by using high-resolution fundus images without compression. Experimental results on the public Messidor data set show that the proposed model can grade DR effectively. Note to Practitioners —This article is motivated by the challenging problem due to the inadequacy of labeled data in medical image analysis and the dispersion of efficient features in high-resolution medical images. As for the inadequacy of labeled data in medical image analysis, the reasons mainly include the followings: 1) the high-quality annotation of medical imaging sample depends heavily on scarce medical expertise which is very expensive and 2) comparing with natural issues, it is more difficult to collect medical images because of privacy issues. It is of great significance to apply deep-learning techniques for diabetic retinopathy (DR) recognition. In this article, the multichannel generative adversarial network (GAN) with semisupervision is developed for DR-aided diagnosis. The proposed model can deal with DR classification problem with inadequacy of labeled data in the following ways: 1) the multichannel generative scheme is proposed to generate a series of subfundus images corresponding to the scattering DR features and 2) the proposed multichannel-based GAN (MGAN) model with semisupervision can make full use of both labeled data and unlabeled data. The experimental results demonstrate that the proposed model outperforms the other representative models in terms of accuracy, area under ROC curve (AUC), sensitivity, and specificity.
74 citations
Additional excerpts
...Besides, the performance of the traditional machine-learning methods often depends on how accurately the features are extracted [10]–[13]....
[...]
TL;DR: The NMS model predicted ankle joint torque best when calibrated with trials during which EMG reached maximum levels, whereas the ANN predicted well when trained with many trials and types of movements, although the ANN prediction may become less reliable when predicting unseen movements.
Abstract: In recent decades, there has been an increasing interest in the use of robotic powered exoskeletons to assist patients with movement disorders in rehabilitation and daily life. Providing assistive torque that compensates for the user’s remaining muscle contributions is a growing and challenging field within exoskeleton control. In this article, ankle joint torques were estimated using electromyography (EMG)-driven neuromusculoskeletal (NMS) model and an artificial neural network (ANN) model in seven movement tasks, including fast walking, slow walking, self-selected speed walking, and isokinetic dorsi/plantar flexion at $60^{\circ }/s$ and $90^{\circ }/s$ . In each method, EMG signals and ankle joint angles were used as input, the models were trained with data from 3-D motion analysis, and ankle joint torques were predicted. Six cases using different motion trials as calibration (for the NMS model)/training (for the ANN) were devised, and the agreement between the predicted and measured ankle joint torques was computed. We found that the NMS model could overall better predict ankle joint torques from EMG and angle data than the ANN model with some exceptions; the ANN predicted ankle joint torques with better agreement when trained with data from the same movement. The NMS model predicted ankle joint torque best when calibrated with trials during which EMG reached maximum levels, whereas the ANN predicted well when trained with many trials and types of movements. In addition, the ANN prediction may become less reliable when predicting unseen movements. Detailed comparative studies of methods to predict ankle joint torque are crucial for determining strategies for exoskeleton control. Note to Practitioners—In exoskeleton control for strength augmentation applied in military, industry, and healthcare applications, providing assistive torque that compensates for the user’s remaining muscle contributions, is a challenging problem. This article predicted the ankle joint torques by electromyography (EMG)-driven neuromusculoskeletal (NMS) model and an artificial neural network (ANN) model in different movements. To the best of our knowledge, this is the first study comparing joint torque prediction performance of EMG-driven model to ANN. In the EMG-driven NMS model, mathematical equations were formulated to reproduce the transformations from EMG signal generation and joint angles to musculotendon forces and joint torques. A three-layer ANN was constructed with an adaptive moment estimation (Adam) optimization method to learn the relationships between the inputs (EMG signals and joint angles) and the outputs (joint torques). In the experiments, we estimated ankle joint torques in gait and isokinetic movements and compared the performance of methods to predict ankle joint torque, relating to how the methods have been calibrated/trained. The detailed analysis of the methods’ performance in predicting ankle joint torque can significantly contribute to determining which model to choose, and under which circumstances, and, thus, be of great benefit for exoskeleton rehabilitation controller design.
34 citations
TL;DR: A new compliance control method of a soft ankle rehabilitation robot that is driven by four PMAs configured in parallel to enable three degrees of freedom movement of the ankle joint is proposed, providing a potentially feasible solution to the patient-in-loop cooperative training strategy.
Abstract: Traditional compliance control of a rehabilitation robot is implemented in task space by using impedance or admittance control algorithms. The soft robot actuated by pneumatic muscle actuators (PMAs) is becoming prominent for patients as it enables the compliance being adjusted in each active link, which, however, has not been reported in the literature. This paper proposes a new compliance control method of a soft ankle rehabilitation robot that is driven by four PMAs configured in parallel to enable three degrees of freedom movement of the ankle joint. A new hierarchical compliance control structure, including a low-level compliance adjustment controller in joint space and a high-level admittance controller in task space, is designed. An adaptive compliance control paradigm is further developed by taking into account patient’s active contribution and movement ability during a previous period of time, in order to provide robot assistance only when it is necessarily required. Experiments on healthy and impaired human subjects were conducted to verify the adaptive hierarchical compliance control scheme. The results show that the robot hierarchical compliance can be online adjusted according to the participant’s assessment. The robot reduces its assistance output when participants contribute more and vice versa, thus providing a potentially feasible solution to the patient-in-loop cooperative training strategy
30 citations
Cites methods from "A Robot-Driven Computational Model ..."
...During the tests, each subject’s active torque and movement ability will be assessed (Zhang et al., 2016) and based on trial and error the control parameters will be determined in experimental manner to ensure that robot can provide specific Frontiers in Neurorobotics | www.frontiersin.org December…...
[...]
References
More filters
TL;DR: OpenSim is developed, a freely available, open-source software system that lets users develop models of musculoskeletal structures and create dynamic simulations of a wide variety of movements to simulate the dynamics of individuals with pathological gait and to explore the biomechanical effects of treatments.
Abstract: Dynamic simulations of movement allow one to study neuromuscular coordination, analyze athletic performance, and estimate internal loading of the musculoskeletal system. Simulations can also be used to identify the sources of pathological movement and establish a scientific basis for treatment planning. We have developed a freely available, open-source software system (OpenSim) that lets users develop models of musculoskeletal structures and create dynamic simulations of a wide variety of movements. We are using this system to simulate the dynamics of individuals with pathological gait and to explore the biomechanical effects of treatments. OpenSim provides a platform on which the biomechanics community can build a library of simulations that can be exchanged, tested, analyzed, and improved through a multi-institutional collaboration. Developing software that enables a concerted effort from many investigators poses technical and sociological challenges. Meeting those challenges will accelerate the discovery of principles that govern movement control and improve treatments for individuals with movement pathologies.
3,621 citations
"A Robot-Driven Computational Model ..." refers methods in this paper
...This ankle model was created based on the lower extremity model developed by Delp et al. [21] in OpenSim [22] and analyzed in MATLAB 2013b....
[...]
...1 and Table I. Muscles included are the same as that in Gait2392 available open through to OpenSim source software [22], but only sections connecting the tibia/fibula and the talus/calcaneus are used, as shown in Fig....
[...]
...Muscles included are the same as that in Gait2392 available open through to OpenSim source software [22], but only sections connecting the tibia/fibula and the talus/calcaneus are used, as shown in Fig....
[...]
...These three rotational axes were modified in OpenSim to orthogonally intersect at a point considered as the rotation center of the ankle–foot complex in robot-assisted therapy....
[...]
...[21] in OpenSim [22] and analyzed in MATLAB 2013b....
[...]
TL;DR: The Standardization and Terminology Committee (STC) of the International Society of Biomechanics proposes definitions of JCS for the ankle, hip, and spine, and suggests that adopting these standards will lead to better communication among researchers and clinicians.
2,650 citations
"A Robot-Driven Computational Model ..." refers methods in this paper
...Definitions of the ankle coordinate system proposed by the Standardization and Terminology Committee of the International Society of Biomechanics were adopted in this model (with dorsiflexion, abduction, and inversion defined as positive, while plantarflexion, adduction, and eversion defined as negative) [27]....
[...]
TL;DR: A model is developed of the human lower extremity to study how changes in musculoskeletal geometry and musculotendon parameters affect muscle force and its moment about the joints and the joint moments calculated with the model compare well with experimentally measured isometric joint moments.
Abstract: A model is developed of the human lower extremity to study how changes in musculoskeletal geometry and musculotendon parameters affect muscle force and its moment about the joints. The lines of action of 43 musculotendon actuators were defined based on their anatomical relationships to three-dimensional bone surface representations. A model for each actuator was formulated to compute its isometric force-length relation. The kinematics of the lower extremity were defined by modeling the hip, knee, ankle, subtalar, and metatarsophalangeal joints. Thus, the force and joint moment that each musculotendon actuator develops can be computed for any body position. The joint moments calculated with the model compare well with experimentally measured isometric joint moments. A graphical interface to the model has also been developed. It allows the user to visualize the musculoskeletal geometry and to manipulate the model parameters to study the biomechanical consequences of orthopaedic surgical procedures. For example, tendon transfer and lengthening procedures can be simulated by adjusting the model parameters according to various surgical techniques. Results of the simulated surgeries can be analyzed quickly in terms of postsurgery muscle forces and other biomechanical variables. >
1,913 citations
"A Robot-Driven Computational Model ..." refers background or methods in this paper
...[21] in OpenSim [22] and analyzed in MATLAB 2013b....
[...]
...The model structure is different with ankle anatomy and published models with ankle joint between the tibia/fibula and the talus and subtalar joint between the talus and the calcaneus [21], [36]....
[...]
TL;DR: It is suggested that, with joint constraints and a global error compensation scheme, the effects of measurement errors on the reconstruction of the musculoskeletal system and subsequent mechanical analyses can be reduced globally.
873 citations
"A Robot-Driven Computational Model ..." refers background in this paper
...The required kinematic inputs to existing computational models are usually obtained from marker-based motion capture systems [35]....
[...]
TL;DR: Accounting for age-related changes in muscle properties, other than decreased maximum isometric force, may be particularly important when simulating movements that require substantial power development.
Abstract: The generation of muscle-actuated simulations that accurately represent the movement of old adults requires a model that accounts for changes in muscle properties that occur with aging. An objective of this study was to adjust the parameters of Hill-type musculo-tendon models to reflect nominal age-related changes in muscle mechanics that have been reported in the literature. A second objective was to determine whether using the parametric adjustments resulted in simulated dynamic ankle torque behavior similar to that seen in healthy old adults. The primary parameter adjustment involved decreasing maximum isometric muscle forces to account for the loss of muscle mass and specific strength with age. A review of the literature suggested the need for other modest adjustments that account for prolonged muscular deactivation, a reduction in maximum contraction velocity, greater passive muscle stiffness and increased normalized force capacity during lengthening contractions. With age-related changes incorporated, a musculo-tendon model was used to simulate isometric and isokinetic contractions of ankle plantarflexor and dorsiflexor muscles. The model predicted that ankle plantarflexion power output during 120 deg/s shortening contractions would be over 40% lower in old adults compared to healthy young adults. These power losses with age exceed the 30% loss in isometric strength assumed in the model but are comparable to 39-44% reductions in ankle power outputs measured in healthy old adults of approximately 70 years of age. Thus, accounting for age-related changes in muscle properties, other than decreased maximum isometric force, may be particularly important when simulating movements that require substantial power development.
565 citations
"A Robot-Driven Computational Model ..." refers background in this paper
...9 and Table IV, which could be caused by various factors like age and gender [38], [39]....
[...]