Abstract The synchronization phenomenon is common to many natural mechanical systems. Joint friction and damping in humans and animals are associated with energy dissipation. A coupled oscillator model is conventionally used to manage multiple joint torque generations to form a limit cycle in an energy dissipation system. The coupling term design and the frequency and phase settings become issues when selecting the oscillator model. The relative coupling relationship between oscillators needs to be predefined for unknown dynamics systems, which is quite challenging problem. We present a simple distributed neural integrators method to induce the limit cycle in unknown energy dissipation systems without using a coupled oscillator. The results demonstrate that synergetic synchronized oscillation could be produced that adapts to different physical environments. Finding the balanced energy injection by neural inputs to form dynamic equilibrium is not a trivial problem, when the dynamics information is not priorly known. The proposed method realized self-organized pattern generation to induce the dynamic equilibrium for different mechanical systems. The oscillation was managed without using the explicit phase or frequency knowledge. However, phase, frequency, and amplitude modulation emerged to form an efficient synchronized limit cycle. This type of distributed neural integrator can be used as a source for regulating multi-joint coordination to induce synergetic oscillations in natural mechanical systems.

<jats:p> Humans can rapidly adapt to new situations, even though they have redundant degrees of freedom (d.f.). Previous studies in neuroscience revealed that human movements could be accounted for by low-dimensional control signals, known as <jats:italic>motor synergies</jats:italic> . Many studies have suggested that humans use the same repertories of motor synergies among similar tasks. However, it has not yet been confirmed whether the combinations of motor synergy repertories can be re-used for new targets in a systematic way. Here we show that the combination of motor synergies can be generalized to new targets that each repertory cannot handle. We use the multi-directional reaching task as an example. We first trained multiple policies with limited ranges of targets by reinforcement learning and extracted sets of motor synergies. Finally, we optimized the activation patterns of sets of motor synergies and demonstrated that combined motor synergy repertories were able to reach new targets that were not achieved with either original policies or single repertories of motor synergies. We believe this is the first study that has succeeded in motor synergy generalization for new targets in new planes, using a full 7-d.f. arm model, which is a realistic mechanical environment for general reaching tasks. </jats:p>

© 2019, © 2019 Informa UK Limited, trading as Taylor & Francis Group and The Robotics Society of Japan. Robot rehabilitation is now recognized as an important method for the efficient recovery. In European Project FP7 BioMot, we have discussed the potential of the robot rehabilitation and proposed the suitable process for it. In this paper, we describe the proposed rehabilitation process and create the theoretical basis for the robot rehabilitation through designing the control system and the patient model. To design the patient model, we describe the source of paralysis and motion controller separately and define the recovery function from the paralysis. In the theoretical analysis of the control system, we show that the robot motions are first adapted to the patient abnormal motions and gradually drive the patient motions to the better ones by the motion support. The singular perturbation analysis proves that the stabilities of the two different process, adaptation to the patient motions and the motion support to the better ones, as a slow motion subsystem and a fast motion subsystem, respectively. The simulation results show that the proposed control system can drive the patients to the better state depending on the patient conditions such as recovery speed and recovery potential. The proposed system can be tuned to fit to the variety of the real patient conditions when we apply it to the real applications.

Functional electrical stimulation (FES) is a neuroprosthetic technique to help restore motor function of spinal cord-injured (SCI) patients. Through delivery of electrical pulses to muscles of motor-impaired subjects, FES is able to artificially induce their muscle contractions. Evoked electromyography (eEMG) is used to record such FES-induced electrical muscle activity and presents a form of M-wave. In order to monitor electrical muscle activity under stimulation and ensure safe stimulation configurations, closed-loop FES control with eEMG feedback is needed to be developed for SCI patients who lose their voluntary muscle contraction ability. This work proposes a closed-loop FES system for real-time control of muscle activation on the triceps surae and tibialis muscle groups through online modulating pulse width (PW) of electrical stimulus. Subject-specific time-variant muscle responses under FES are explicitly reflected by muscle excitation model, which is described by Hammerstein system with its input and output being, respectively, PW and eEMG. Model predictive control is adopted to compute the PW based on muscle excitation model which can online update its parameters. Four muscle activation patterns are provided as desired control references to validate the proposed closed-loop FES control paradigm. Real-time experimental results on three able-bodied subjects and five SCI patients in clinical environment show promising performances of tracking the aforementioned reference muscle activation patterns based on the proposed closed-loop FES control scheme.

Copyright © 2018 Okajima, Tournier, Alnajjar, Hayashibe, Hasegawa and Shimoda. An important function missing from current robotic systems is a human-like method for creating behavior from symbolized information. This function could be used to assess the extent to which robotic behavior is human-like because it distinguishes human motion from that of human-made machines created using currently available techniques. The purpose of this research is to clarify the mechanisms that generate automatic motor commands to achieve symbolized behavior. We design a controller with a learning method called tacit learning, which considers system-environment interactions, and a transfer method called mechanical resonance mode, which transfers the control signals into a mechanical resonance mode space (MRM-space). We conduct simulations and experiments that involve standing balance control against disturbances with a two-degree-of-freedom inverted pendulum and bipedal walking control with humanoid robots. In the simulations and experiments on standing balance control, the pendulum can become upright after a disturbance by adjusting a few signals in MRM-space with tacit learning. In the simulations and experiments on bipedal walking control, the robots realize a wide variety of walking by manually adjusting a few signals in MRM-space. The results show that transferring the signals to an appropriate control space is the key process for reducing the complexity of the signals from the environment and achieving diverse behavior.

Poststroke hemiplegic patients often show altered weight distribution with balance disorders, increasing their risk of fall. Conventional balance training, though powerful, suffers from scarcity of trained therapists, frequent visits to clinics to get therapy, one-on-one therapy sessions, and monotony of repetitive exercise tasks. Thus, technology-assisted balance rehabilitation can be an alternative solution. Here, we chose virtual reality as a technology-based platform to develop motivating balance tasks. This platform was augmented with off-the-shelf available sensors such as Nintendo Wii balance board and Kinect to estimate one's center of mass (CoM). The virtual reality-based CoM-assisted balance tasks (Virtual CoMBaT) was designed to be adaptive to one's individualized weight-shifting capability quantified through CoM displacement. Participants were asked to interact with Virtual CoMBaT that offered tasks of varying challenge levels while adhering to ankle strategy for weight shifting. To facilitate the patients to use ankle strategy during weight-shifting, we designed a heel lift detection module. A usability study was carried out with 12 hemiplegic patients. Results indicate the potential of our system to contribute to improving one's overall performance in balance-related tasks belonging to different difficulty levels.

Copyright © 2018 Okajima, Tournier, Alnajjar, Hayashibe, Hasegawa and Shimoda. An important function missing from current robotic systems is a human-like method for creating behavior from symbolized information. This function could be used to assess the extent to which robotic behavior is human-like because it distinguishes human motion from that of human-made machines created using currently available techniques. The purpose of this research is to clarify the mechanisms that generate automatic motor commands to achieve symbolized behavior. We design a controller with a learning method called tacit learning, which considers system-environment interactions, and a transfer method called mechanical resonance mode, which transfers the control signals into a mechanical resonance mode space (MRM-space). We conduct simulations and experiments that involve standing balance control against disturbances with a two-degree-of-freedom inverted pendulum and bipedal walking control with humanoid robots. In the simulations and experiments on standing balance control, the pendulum can become upright after a disturbance by adjusting a few signals in MRM-space with tacit learning. In the simulations and experiments on bipedal walking control, the robots realize a wide variety of walking by manually adjusting a few signals in MRM-space. The results show that transferring the signals to an appropriate control space is the key process for reducing the complexity of the signals from the environment and achieving diverse behavior.

BackgroundFatigue induced hand tremor (FIT) is a primary limiting concern for the prolonged surgical intervention in minimally invasive surgery (MIS) and robot-assisted-minimally invasive surgery (RAMIS). A thorough analysis is necessary to understand the FIT characteristics in laparoscopic tool movement. The primary aim of this study is to perform a differential analysis of the elbow and wrist tremor due to muscle fatigue in laparoscopic manoeuvring. MethodsWe have introduced a joint angle based tremor analysis method, which enables us to perform a differential study of FIT characteristics at the individual joint. Experimental data was acquired from a group of subjects during static and dynamic laparoscopic movement in an imitative RAMIS master manipulation scenario. A repetitive task was performed with a total span of 1h for observing the effect of muscle fatigue. Along with the joint angle variation, surface electromyography (sEMG) signal was also studied in the analysis. ResultsThe wrist tremor is more predominant than tremor generated at the elbow, especially in highly fatigued condition. The high-frequency tremor (&gt;4Hz) is contributed by the wrist joint. Moreover, the variation of the wrist and elbow tremor ratio was found to be dependent upon the experience of the surgeons. ConclusionsIn this work, we have investigated the attribution of elbow and wrist joints in FIT during laparoscopic tool manipulation. The outcomes may be useful for the design of robot-assisted surgical manipulator, and can be used for quality assessment of surgical training as well.

Performance assessment of human movement is critical in diagnosis and motor-control rehabilitation. Recent developments in portable sensor technology enable clinicians to measure spatiotemporal aspects to aid in the neurological assessment. However, the extraction of quantitative information from such measurements is usually done manually through visual inspection. This paper presents a novel framework for automatic human movement assessment that executes segmentation and motor performance parameter extraction in time-series of measurements from a sequence of human movements. We use the elements of a Switching Linear Dynamic System model as building blocks to translate formal definitions and procedures from human movement analysis. Our approach provides a method for users with no expertise in signal processing to create models for movements using labeled dataset and later use it for automatic assessment. We validated our framework on preliminary tests involving six healthy adult subjects that executed common movements in functional tests and rehabilitation exercise sessions, such as sit-to-stand and lateral elevation of the arms and five elderly subjects, two of which with limited mobility, that executed the sit-to-stand movement. The proposed method worked on random motion sequences for the dual purpose of movement segmentation (accuracy of 72%-100%) and motor performance assessment (mean error of 0%-12%).

Reliable detection of error from electroencephalography (EEG) signals as feedback while performing a discrete target selection task across sessions and subjects has a huge scope in real-time rehabilitative application of Brain-computer Interfacing (BCI). Error Related Potentials (ErrP) are EEG signals which occur when the participant observes an erroneous feedback from the system. ErrP holds significance in such closed-loop system, as BCI is prone to error and we need an effective method of systematic error detection as feedback for correction. In this paper, we have proposed a novel scheme for online detection of error feedback directly from the EEG signal in a transferable environment (i.e., across sessions and across subjects). For this purpose, we have used a P300-speller dataset available on a BCI competition website. The task involves the subject to select a letter of a word which is followed by a feedback period. The feedback period displays the letter selected and, if the selection is wrong, the subject perceives it by the generation of ErrP signal. Our proposed system is designed to detect ErrP present in the EEG from new independent datasets, not involved in its training. Thus, the decoder is trained using EEG features of 16 subjects for single-trial classification and tested on 10 independent subjects. The decoder designed for this task is an ensemble of linear discriminant analysis, quadratic discriminant analysis, and logistic regression classifier. The performance of the decoder is evaluated using accuracy, F1-score, and Area Under the Curve metric and the results obtained is 73.97, 83.53, and 73.18%, respectively.

Fatigue induced hand tremor (FIT) is an unavoidable phenomenon, which substantially limits the accuracy of the surgical manipulation for long duration laparoscopic surgeries. Filtering intended motion from tremor is a challenging task as the properties of tremor change with increasing muscle fatigue levels. Muscle fatigue induced hand tremor has highly nonlinear and nonstationary characteristics that need a filtering strategy different from the conventional filters. Empirical Mode Decomposition (EMD) based filters have become popular in the recent past for its enhanced nonlinear signal handling capability. EMD based filtering strategy is case specific in nature as the EMD does not have any general analytical formulation unlike other (Kernel based) popular filtering techniques. In this work, we have addressed the tremor filtering issue with the help of EMD and the probability distribution characteristics analysis of Intrinsic Mode Functions (IMF) of the tremulous laparoscopic tool trajectory. A modified distribution asymmetry measure was employed to find out the threshold IMF for reconstruction of tremor free motion at different fatigue levels. In order to find the robustness of the proposed technique, the compensation strategy has been tested extensively on synthetic signal and experimentally acquired signals. Filtering threshold at different fatigue levels was also demonstrated for various subjects. Despite the time-varying properties of tremor, the proposed filtering strategy substantiates its efficacy to diminish the effect of tremor which was not possible by the conventional fixed cut-off filtering techniques. (C) 2016 Elsevier Ltd. All rights reserved.

Background: Transcranial direct current stimulation (tDCS) has been shown to perturb both cortical neural activity and hemodynamics during (online) and after the stimulation, however mechanisms of these tDCS-induced online and after-effects are not known. Here, online resting-state spontaneous brain activation may be relevant to monitor tDCS neuromodulatory effects that can be measured using electroencephalography (EEG) in conjunction with near-infrared spectroscopy (NIRS). Method: We present a Kalman Filter based online parameter estimation of an autoregressive (ARX) model to track the transient coupling relation between the changes in EEG power spectrum and NIRS signals during anodal tDCS (2 mA, 10 min) using a 4 x 1 ring high-definition montage. Results: Our online ARX parameter estimation technique using the cross-correlation between log (base 10) transformed EEG band-power (0.5-11.25 Hz) and NIRS oxy-hemoglobin signal in the low frequency (&lt;= 0.1 Hz) range was shown in 5 healthy subjects to be sensitive to detect transient EEG-NIRS coupling changes in resting-state spontaneous brain activation during anodal tDCS. Conventional sliding window cross-correlation calculations suffer a fundamental problem in computing the phase relationship as the signal in the window is considered time-invariant and the choice of the window length and step size are subjective. Here, Kalman Filter based method allowed online ARX parameter estimation using time varying signals that could capture transients in the coupling relationship between EEG and NIRS signals. Conclusion: Our new online ARX model based tracking method allows continuous assessment of the transient coupling between the electrophysiological (EEG) and the hemodynamic (NIRS) signals representing resting-state spontaneous brain activation during anodal tDCS. Published by Elsevier B.V.

This paper proposes a novel brain-machine interfacing (BMI) paradigm for control of a multijoint redundant robot system. Here, the user would determine the direction of end-point movement of a 3-degrees of freedom (DOF) robot arm using motor imagery electroencephalography signal with co-Adaptive decoder (adaptivity between the user and the decoder) while a synergetic motor learning algorithm manages a peripheral redundancy in multi-DOF joints toward energy optimality through tacit learning. As in human motor control, torque control paradigm is employed for a robot to be adaptive to the given physical environment. The dynamic condition of the robot arm is taken into consideration by the learning algorithm. Thus, the user needs to only think about the end-point movement of the robot arm, which allows simultaneous multijoints control by BMI. The support vector machine-based decoder designed in this paper is adaptive to the changing mental state of the user. Online experiments reveals that the users successfully reach their targets with an average decoder accuracy of over 75% in different end-point load conditions.

This paper proposes a novel brain-machine interfacing (BMI) paradigm for control of a multijoint redundant robot system. Here, the user would determine the direction of end-point movement of a 3-degrees of freedom (DOF) robot arm using motor imagery electroencephalography signal with co-adaptive decoder (adaptivity between the user and the decoder) while a synergetic motor learning algorithm manages a peripheral redundancy in multi-DOF joints toward energy optimality through tacit learning. As in human motor control, torque control paradigm is employed for a robot to be adaptive to the given physical environment. The dynamic condition of the robot arm is taken into consideration by the learning algorithm. Thus, the user needs to only think about the end-point movement of the robot arm, which allows simultaneous multijoints control by BMI. The support vector machine-based decoder designed in this paper is adaptive to the changing mental state of the user. Online experiments reveals that the users successfully reach their targets with an average decoder accuracy of over 75% in different end-point load conditions.

Background: Functional electrical stimulation (FES) is a neuroprosthetic technique for restoring lost motor function of spinal cord injured (SCI) patients and motor-impaired subjects by delivering short electrical pulses to their paralyzed muscles or motor nerves. FES induces action potentials respectively on muscles or nerves so that muscle activity can be characterized by the synchronous recruitment of motor units with its compound electromyography (EMG) signal is called M-wave. The recorded evoked EMG (eEMG) can be employed to predict the resultant joint torque, and modeling of FES-induced joint torque based on eEMG is an essential step to provide necessary prediction of the expected muscle response before achieving accurate joint torque control by FES. Methods: Previous works on FES-induced torque tracking issues were mainly based on offline analysis. However, toward personalized clinical rehabilitation applications, real-time FES systems are essentially required considering the subject-specific muscle responses against electrical stimulation. This paper proposes a wireless portable stimulator used for estimating/predicting joint torque based on real time processing of eEMG. Kalman filter and recurrent neural network (RNN) are embedded into the real-time FES system for identification and estimation. Results: Prediction results on 3 able-bodied subjects and 3 SCI patients demonstrate promising performances. As estimators, both Kalman filter and RNN approaches show clinically feasible results on estimation/prediction of joint torque with eEMG signals only, moreover RNN requires less computational requirement. Conclusion: The proposed real-time FES system establishes a platform for estimating and assessing the mechanical output, the electromyographic recordings and associated models. It will contribute to open a new modality for personalized portable neuroprosthetic control toward consolidated personal healthcare for motor-impaired patients.

Functional Electrical Stimulation (FES) stimulates the affected region of the human body thus providing a neuroprosthetic interface to non-recovered muscle groups. FES in combination with Brain-computer interfacing (BCI) has a wide scope in rehabilitation because this system can directly link the cerebral motor intention of the users with its corresponding peripheral mucle activations. Such a rehabilitative system would contribute to improve the cortical and peripheral learning and thus, improve the recovery time of the patients. In this paper, we examine the effect of electrical stimulation by FES on the electroencephalography (EEG) during learning of a motor imagery task. The subjects are asked to perform four motor imagery tasks over six sessions and the features from the EEG are extracted using common spatial algorithm and decoded using linear discriminant analysis classifier. Feedback is provided in form of a visual medium and electrical stimulation representing the distance of the features from the hyperplane. Results suggest a significant improvement in the classification accuracy when the subject was induced with electrical stimulation along with visual feedback as compared to the standard visual one.

As the center of mass (CoM) position can be used to determine stability, current rehabilitation standards may be improved by tracking it. A personalized CoM estimate can be obtained using the statically equivalent serial chain (SESC) once the model parameters are identified. The identification phase can be completed using low-cost sensors (Kinect and Wii balance board) outside the laboratory making CoM estimation feasible in a patient's home. This paper focuses on: 1) improving the SESC identification quality and speed and 2) using the estimated CoM to determine stability. Identification time is reduced by creating a visual adaptive interface where the subject's limbs are colored based on the convergence of the SESC parameters. A study was conducted on eight subjects and showed a faster convergence and lower root mean square error (RMSE) when the adaptive interface was used. We found that a model capable of estimating the CoM position with an RMSE of 27 mm could be obtained after only 90 s of identification when the interface was used, whereas twice as much time was needed when the interface was not used. The interface that was developed can be used by a subject to track his/ her CoM position in a self-directed way. Stability was determined for a squat task using a dynamic index obtained from the estimated CoM trajectory and using only Kinect measurements. This shows one potential application for home rehabilitation and monitoring.

We investigated the synthesis of electrical stimulation patterns for functional movement restoration in human paralyzed limbs. We considered the knee joint system, co-activated by the stimulated quadriceps and hamstring muscles. This synthesis is based on optimized functional electrical stimulation (FES) patterns to minimize muscular energy consumption and movement efficiency criteria. This two-part work includes a multi-scale physiological muscle model, based on Huxley's formulation. In the simulation, three synthesis strategies were investigated and compared in terms of muscular energy consumption and co-contraction levels. In the experimental validation, the synthesized FES patterns were carried out on the quadriceps-knee joint system of four complete spinal cord injured subjects. Surface stimulation was applied to all subjects, except for one FES-implanted subject who received neural stimulation. In each experimental validation, the model was adapted to the subject through a parameter identification procedure. Simulation results were successful and showed high co-contraction levels when reference trajectories were tracked. Experimental validation results were encouraging, as the desired and measured trajectories showed good agreement, with an 8.4 % rms error in a subject without substantial time-varying behavior. We updated the maximal isometric force in the model to account for time-varying behavior, which improved the average rms errors from 31.4 to 13.9 % for all subjects.

Control of robotic joints movements requires the generation of appropriate torque and force patterns, coordinating the kinematically and dynamically complex multijoints systems. Control theory coupled with inverse and forward internal models are commonly used to map a desired endpoint trajectory into suitable force patterns. In this paper, we propose the use of tacit learning to successfully achieve similar tasks without using any kinematic model of the robotic system to be controlled. Our objective is to design a new control strategy that can achieve levels of adaptability similar to those observed in living organisms and be plausible from a neural control viewpoint. If the neural mechanisms used for mapping goals expressed in the task-space into control-space related command without using internal models remain largely unknown, many neural systems rely on data accumulation. The presented controller does not use any internal model and incorporates knowledge expressed in the task space using only the accumulation of data. Tested on a simulated two-link robot system, the controller showed flexibility by developing and updating its parameters through learning. This controller reduces the gap between reflexive motion based on simple accumulation of data and execution of voluntarily planned actions in a simple manner that does not require complex analysis of the dynamics of the system.

The human brain can be considered as a graphical network having different regions with specific functionality and it can be said that a virtual functional connectivity are present between these regions. These regions are regarded as nodes and the functional links are regarded as the edges between them. The intensity of these functional links depend on the activation of the lobes while performing a specific task(e.g. motor imagery tasks, cognitive tasks and likewise). The main aim of this study is to understand the activation of the parts of the brain while performing three types of motor imagery tasks with the help of graph theory. Two indices of the graph, namely Network Density and Node Strength are calculated for 32 electrodes placed on the subject's head covering all the brain lobes and the nodes having higher intensity are identified.

Human motion assessment is key for motor-control rehabilitation. Using standardized definitions and spatio-temporal features - usually presented as a movement cycle diagram- specialists can associate kinematic measures to progress in rehabilitation therapy or motor impairment due to trauma or disease. Although devices for capturing human motion today are cheap and widespread, the automatic interpretation of kinematic data for rehabilitation is still poor in terms of quantitative performance evaluation. In this paper we present an automatic approach to extract spatiotemporal features from kinematic data and present it as a cycle diagram. This is done by translating standard definitions from human movement analysis into mathematical elements of a Switching Linear Dynamic System model. The result is a straight-forward procedure to learn a tracking model from a sample execution. This model is robust when used to automatically extract the movement cycle diagram of the same motion (the Sit-Stand-Sit, as an example) executed in different subject-specific manner such as his own motion speed.

Accurate and real time balance estimation can be used to improve home based rehabilitation systems. We developed a personalized balance measurement, making use of the subject-specific center of mass (CoM) estimate offered by the statically equivalent serial chain (SESC) method and the zero rate of change of angular momentum (ZRAM) concept to evaluate balance during a series of dynamic motions. Two healthy subjects were asked to stand on a Wii balance board and their SESC parameters were identified. A set of dynamic motions was then recorded and the rate of change of centroidal angular momentum and the distance of the ZRAM point to the center line of the support polygon were obtained. A good match between both metrics was found. Additionally, we developed a real time application based on Kinect measurements that determines the ZRAM position, in real time, and displays it to the subject in the form of visual feedback. In this way the ZRAM can be used to evaluate balance in home-rehabilitation for any motion.

This study investigates the possibility of using the so-called Statically Equivalent Serial Chain approach to estimate the subject-specific 3D whole-body centre of mass (CoM) location. This approach is based on a compact formulation of the 3D whole-body CoM position associated with a least squares identification process. This process requires a calibration phase that uses stereophotogrammetric and dynamometric data collected in selected static postures. After this calibration phase, the instantaneous position of the identified subject-specific 3D whole-body CoM can be estimated for any motor task using kinematic data only. This approach was experimentally validated on twelve healthy young subjects. The Statically Equivalent Serial Chain solution was validated during static trials with the centre of pressure, with the double integrated ground reaction forces during dynamic tasks, and also compared with a segmental method using a stereophotogrammetric system and anthropometric tables. Considerations relative to the choice of algorithm parameters, such as the number of necessary static postures and their time duration, are discussed. The proposed method shows much smaller differences between the projection of the centre of mass and the centre of pressure (root mean square value under 3.5%) than the method using anthropometric tables (root mean square value over 9%). Same conclusion can be made during dynamic tasks with a smaller difference obtained for SESC (root mean square value under 4% at contrary the 20% obtained with anthropometric table). (C) 2014 Elsevier B.V. All rights reserved.

Functional electrical stimulation (FES) is a useful technique for restoring motor functions for spinal cord injured (SCI) patients. Muscle contractions can be artificially driven through delivery of electrical pulses to impaired muscles, and the electrical activity of contracted muscles under stimulus recorded by electromyography (EMG) is called M-wave. The FES-induced muscle activation which is represented by evoked EMG recordings can indicate the muscle state. Accurate control of muscle activation level by FES is the preliminary step for achieving more complicated FES control tasks. This paper proposes a real-time FES system for control of muscle activation by online modulating pulse width of stimulus. The excitation muscle dynamics is modelled by Hammerstain system with stimulus pulse width and eEMG as input and output respectively. The model predictive control strategy is adopted to compute the pulse width command sent to the Vivaltis wireless stimulator. Four reference muscle activation patterns are provided to test and validate the real-time closed-loop FES control system. Real-time control results on one able-bodied subject show promising control performances.

Human movement is produced resulting from synergetic combinations of multiple muscle contractions. The resultant joint movement can be estimated through the related multiple-muscle activities, which is formulated as the forward problem. Neuroprosthetic applications may benefit from cocontraction of agonist and antagonist muscle pairs to achieve more stable and robust joint movements. It is necessary to estimate the activation of each individual muscle from desired joint torque(s), which is the inverse problem. A synergy-based solution is presented for the inverse estimation of multiple muscle activations from joint movement, focusing on one degree-of-freedom tasks. The approach comprises muscle synergy extraction via the nonnegative matrix factorization algorithm. Cross validation is performed to evaluate the method for prediction accuracy based on experimental data from ten able-bodied subjects. The results demonstrate that the approach succeeds to inversely estimate the multiple muscle activities from the given joint torque sequence. In addition, the other one's averaged synergy ratio was applied for muscle activation estimation with leave-one-out cross-validation manner, which resulted in 9.3% estimation error over all the subjects. The obtained results support the common muscle synergy-based neuroprosthetics control concept.

Slow-growing, infiltrative brain tumours may modify the electrophysiological balance between the two hemispheres. To determine whether and how asymmetry in interhemispheric excitability might occur following "wide-awake surgery" for this type of tumour, we recorded electroencephalograms during a simple visuo-manual reaction time paradigm performed by five patients between 3 and 12 months after surgery. Interhemispheric excitability asymmetries were computed by comparing the amplitudes of event-related potentials (ERPs) in the injured hemisphere to those in the healthy hemisphere. For the two patients with the smallest lesions (7.1 and 11.5 cm(3), respectively), increased excitability within the ipsilesional hemisphere was evidenced by characteristics increases in the ERP amplitude at several sites, with few occurrences in the contralesional hemisphere. For smaller lesions (and under certain experimental conditions), cortical excitability in the injured hemisphere may increase in order to maintain local compensation. In addition, we observed and increased excitability in the contralesional frontal homologue for one patient who underwent an extensive resection. Post-operative monitoring of interhemispheric asymmetries in ERP amplitudes is of value for determining task constraints inducing electrophysiological imbalance and guiding rehabilitation.

Introduction: The purpose of this study was to propose a method that allows extraction of the current muscle state under electrically induced fatigue. Methods: The triceps surae muscle of 5 subjects paralyzed by spinal cord injury was fatigued by intermittent electrical stimulation (5 X 5 trains at 30 HZ). Classical fatigue indices representing muscle contractile properties [peak twitch (Pt) and half-relaxation time (HRT)] were assessed before and after each 5-train series and were used to identify 2 relevant parameters (F-m, U-r) of a previously developed mathematical model using the Sigma-Point Kalman Filter. Results: Pt declined significantly during the protocol, whereas HRT remained unchanged. Identification of the model parameters with experimental data yielded a model-based fatigue assessment that gave a more stable evaluation of fatigue than classical parameters. Conclusions: This work reinforces clinical research by providing a tool that clinicians can use to monitor fatigue development during stimulation.

The trajectory of the whole body center of mass (CoM) is useful as a reliable metric of postural stability. If the evaluation of a subject-specific CoM were available outside of the laboratory environment, it would improve the assessment of the effects of physical rehabilitation. This paper develops a method that enables tracking CoM position using low-cost sensors that can be moved around by a therapist or easily installed inside a patient's home. Here, we compare the accuracy of a personalized CoM estimation using the statically equivalent serial chain (SESC) method and measurements obtained with the Kinect to the case of a SESC obtained with high-end equipment (Vicon). We also compare these estimates to literature-based ones for both sensors. The method was validated with seven able-bodied volunteers for whom the SESC was identified using 40 static postures. The literature-based estimation with Vicon measurements had a average error 24.9 +/- 3.7 mm; this error was reduced to 12.8 +/- 9.1 mm with the SESC identification. When using Kinect measurements, the literature-based estimate had an error of 118.4 +/- 50.0 mm, while the SESC error was 26.6 +/- 6.0 mm. The subject-specific SESC estimate using low-cost sensors has an equivalent performance as the literature-based one with high-end sensors. The SESC method can improve CoM estimation of elderly and neurologically impaired subjects by considering variations in their mass distribution.

One of the challenging issues in computational rehabilitation is that there is a large variety of patient situations depending on the type of neurological disorder. Human characteristics are basically subject specific and time variant; for instance, neuromuscular dynamics may vary due to muscle fatigue. To tackle such patient specificity and time-varying characteristics, a robust bio-signal processing and a precise model-based control which can manage the nonlinearity and time variance of the system, would bring break-through and new modality toward computational intelligence (CI) based rehabilitation technology and personalized neuroprosthetics. Functional electrical stimulation (FES) is a useful technique to assist restoring motor capability of spinal cord injured (SCI) patients by delivering electrical pulses to paralyzed muscles. However, muscle fatigue constraints the application of FES as it results in the time-variant muscle response. To perform adaptive closed-loop FES control with actual muscle response feedback taken into account, muscular torque is essential to be estimated accurately. However, inadequacy of the implantable torque sensor limits the direct measurement of the time-variant torque at the joint. This motivates the development of methods to estimate muscle torque from bio-signals that can be measured. Evoked electromyogram (eEMG) has been found to be highly correlated with FES-induced torque under various muscle conditions, indicating that it can be used for torque/force prediction. A nonlinear ARX (NARX) type model is preferred to track the relationship between eEMG and stimulated muscular torque. This paper presents a NARX recurrent neural network (NARX-RNN) model for identification/prediction of FES-induced muscular dynamics with eEMG. The NARX-RNN model may possess novelty of robust prediction performance. Due to the difficulty of choosing a proper forgetting factor of Kalman filter for predicting time-variant torque with eEMG, the presented NARX-RNN could be considered as an alternative muscular torque predictor. Data collected from five SCI patients is used to evaluate the proposed NARX-RNN model, and the results show promising estimation performances. In addition, the general importance regarding CI-based motor function modeling is introduced along with its potential impact in the rehabilitation domain. The issue toward personalized neuroprosthetics is discussed in detail with the potential role of CI-based identification and the benefit for motor-impaired patient community.

A human motor system can improve its behavior toward optimal movement. The skeletal system has more degrees of freedom than the task dimensions, which incurs an ill-posed problem. The multijoint system involves complex interaction torques between joints. To produce optimal motion in terms of energy consumption, the so-called cost function based optimization has been commonly used in previous works. Even if it is a fact that an optimal motor pattern is employed phenomenologically, there is no evidence that shows the existence of a physiological process that is similar to such a mathematical optimization in our central nervous system. In this study, we aim to find a more primitive computational mechanism with a modular configuration to realize adaptability and optimality without prior knowledge of system dynamics. We propose a novel motor control paradigm based on tacit learning with task space feedback. The motor command accumulation during repetitive environmental interactions, play a major role in the learning process. It is applied to a vertical cyclic reaching which involves complex interaction torques. We evaluated whether the proposed paradigm can learn how to optimize solutions with a 3-joint, planar biomechanical model. The results demonstrate that the proposed method was valid for acquiring motor synergy and resulted in energy efficient solutions for different load conditions. The case in feedback control is largely affected by the interaction torques. In contrast, the trajectory is corrected over time with tacit learning toward optimal solutions. Energy efficient solutions were obtained by the emergence of motor synergy. During learning, the contribution from feedforward controller is augmented and the one from the feedback controller is significantly minimized down to 12% for no load at hand, 16% for a 0.5 kg load condition. The proposed paradigm could provide an optimization process in redundant system with dynamic-model-free and cost-function-free approach.

Functional electrical stimulation (FES) is a useful rehabilitation technique for restoring motor capability of spinal cord injured (SCI) patients by artificially driving muscle contractions. Real-time FES systems with online modulation ability are in great need for clinical applications. In this work, a system for real-time estimation of joint torque with evoked electromyography (eEMG) is presented. Kalman filter (KF) is adopted and embedded into the system as the online torque estimator. The real-time estimation system would be promising toward FES control with consideration of torque changes caused by muscle fatigue.

This paper presents a novel approach for simulating 3D muscle deformations with complex architectures. The approach consists in choosing the best model formulation in terms of computation cost and accuracy, that mixes a volumetric tissue model based on finite element method (3D FEM), a muscle fiber model (Hill contractile 1D element) and a membrane model accounting for aponeurosis tissue (2D FEM). The separate models are mechanically binded using barycentric embeddings. Our approach allows the computation of several fiber directions in one coarse finite element, and thus, strongly decreases the required finite element resolution to predict muscle deformation during contraction. Using surface registration, fibers tracks of specific architecture can be transferred from a template to subject morphology, and then simulated. As a case study, three different architectures are simulated and compared to their equivalent one dimensional Hill wire model simulations.

Accuracy of laparoscopic surgery gets affected by the hand tremor of the surgeons. Though cognitive load is inevitable in such activity which promotes tremor, muscle fatigue induced tremor is significant among the most important sources of tremor. Characteristic of fatigue induced hand tremor and its dominant directional properties are reported in this work. For a fixed laparoscopic tool grip with temporally synchronized predefined task protocols, characteristics of fatigue induced tremors have been studied. Dominant component of tremor was found to be in the sagittal plane in case of both static and dynamic tasks. In order to relate it with the muscle fatigue level, spectral properties of surface electromyography (SEMG) were also investigated simultaneously. A study of transient effect on tool positioning was also included, which conjointly advocates the other experimental results on fatigue induced hand tremor as well.

Living organisms are characterized by their smooth adaptability to environmental changes and their robustness against morphological modifications. To investigate the computational mechanisms behind such learning scheme, we proposed tacit learning as a novel learning method. In tacit learning, there are no clear distinctions between learning and motor control: learning is a simple accumulation process embedded in the controller. In previous work, tacit learning was applied with success to bipedal locomotion of a 36 DoF humanoid robot. In this paper, we generalize the structure of the controller such as applying adaptive integration to a wider range of systems and behaviors. This is achieved by applying the principle of tacit learning in a hierarchical fashion, in which the value of a virtual periodic dynamic variable is tuned for continuous adaptation. This resulting PD-PI (proportional-derivative periodic-integration) controller preserves the advantages of tacit learning that the controllers do not include any prior knowledge of the system in which they are embedded. It also shares with biological systems the property that control and adaptation progress without explicit distinction between them.

This paper proposed a closed-loop torque control strategy of functional electrical stimulation (FES) with the aim of obtaining an accurate, safe, and robust FES system. Generally, FES control systems are faced with the challenge of how to deal with time-variant muscle dynamics due to physiological and biochemical factors (such as fatigue). The degraded muscle force needs to be compensated in order to ensure the accuracy of the motion restored by FES. Another challenge concerns the fact that implantable sensors are unavailable to feedback torque information for FES in humans. As FES-evoked electromyography (EMG) represents the activity of stimulated muscles, and also enables joint torque prediction as presented in our previous studies, here we propose an EMG-feedback predictive controller of FES to control joint torque adaptively. EMG feedback contributes to taking the activated muscle state in the FES torque control system into account. The nature of the predictive controller facilitates prediction of the muscle mechanical response and the system can therefore control joint torque from EMG feedback and also respond to time-variant muscle state changes. The control performance, fatigue compensation and aggressive control suppression capabilities of the proposed controller were evaluated and discussed through experimental and simulation studies.

Center of mass (CoM) estimation can be used to evaluate human stability during rehabilitation. A personalized estimation can be obtained using the serial equivalent static chain (SESC) method, calibrated using a series of static postures. The estimation accuracy is dependent on the number and quality of poses used during calibration. Currently, this limits the method's application to unimpaired individuals. We present a preliminary study of a SESC identified in a multi-contact scenario during a Sit-to-Stand task. Stanford's SAI (Simulation and Active Interface) platform was used to emulate motion and predict relevant reaction forces. The CoM estimation obtained is valid for motions similar to those used during identification. Using a three-dimensional model, the estimated mean error was less than 26 millimetres for a Sit-to-Stand task involving displacements along all axes. As such, personalized CoM estimation can be available for patients with a limited range of whole body motion.

A human's center of mass (CoM) trajectory is useful to evaluate the dynamic stability during daily life activities such as walking and standing up. To estimate the subject-specific CoM position in the home environment, we make use of a statically equivalent serial chain (SESC) developed with a portable measurement system. In this paper we implement a constrained Kalman filter to achieve an online estimation of the SESC parameters while accounting for the human body's bilateral symmetry. This results in constraining SESC parameters to be consistent with the human skeletal model used. The proposed identification method can inform the subject or the therapist, in real-time, about the quality of the on-going CoM estimation. This information can be helpful to reduce the identification time and establish a personalized protocol. A Kinect is used as a markerless motion capture system for measuring limb orientations while the Wii board is used to measure the subject's center of pressure (CoP) during the identification phase. CoP measurements and Kinect data were recorded for four able-bodied subjects. The recorded data was then given to the proposed recursive algorithm to identify the parameters of the SESC online. A cross-validation test was performed to verify the identification performance. The results for these subjects are shown and discussed.

The signal measured with an electromyogram (EMG) is the summation of all action potentials of motor units active at a certain time. According to previous literature, one can establish the relationship between torque and EMG/activations in a forward way, i.e., employing EMG of multiple channels to estimate the joint torque. Once the relationship is established, the torque can be predicted with EMG recordings. However, in some applications of neuroprosthetics where we need to make muscle control, it is required to inversely have an insight regarding the muscle activations under a specific motion scenario from the corresponding torque. Motivated by this point, this paper investigates inverse estimation of muscle activations in random contractions at the ankle joint. Local multiple regression is exploited for finding the relationship between muscle activations and torque. Such technique is able to rebuild the relationship between muscle activations and joint torque inversely based on experimental data obtained from five able-bodied subjects, and the resultant optimal weight matrix can indicate each muscle's contribution in the production of the torque. Further cross validation on prediction of muscle activations with joint torque with optimal weights shows that such approach may possess promising performance.

The dynamics of multijoint limbs often causes complex dynamic interaction torques which are the inertial effect of other joints motion. It is known that Cerebellum takes important role in a motor learning by developing the internal model. In this paper, we propose a novel computational control paradigm in vertical reaching task which involves the management of interaction torques and gravitational effect. The obtained results demonstrate that the proposed method is valid for acquiring motor synergy in the system with actuation redundancy and resulted in the energy efficient solutions. It is highlighted that the tacit learning in vertical reaching task can bring computational adaptability and optimality with model-free and cost-function-free approach differently from previous studies.

In order to understand the human motion control strategies and to restore these functions, or to artificially generate limbs motion it is necessary to have an accurate understanding of the limb dynamics. The inertial parameters can be identify easily, however the joint dynamics is still difficult to model due to the time change with muscle contraction level, fatigue and non-linear dynamics. Using Functional Electrical Stimulation (FES) we propose to identify the joint active dynamics with the pendulum test and to establish a relationship between the level of muscle contraction induced by the stimulation and the visco-elasticity. We measure the data of 2 healthy subjects and propose a model for the knee joint visco-elasticity changes. © 2012 IEEE.

Computer simulation is promising numerical tool to study muscle volumetric deformations. However, most models are facing very long computation time and thus are based on simplified wire Hill muscle model. The purpose of this study is to develop a real-time three-dimensional biomechanical model of volumetric muscle based on modified Hill model for the active stress which is controlled from EMG recordings. Finite element model is used to estimate the passive behavior of the muscle and tendons during contraction. We demonstrate that this 3D model implementation is very cost effective with respect to the computation time and the simulation gives good results compared to real measured data. Thus, this effective implementation will allow implementing much more complex and realistic models considering the muscle as volumetric continuum, with moderate computation time.

Functional electrical stimulation (FES) is able to restore motor function of spinal cord injured (SCI) patients. To make adaptive FES control taking into account the actual muscle state with muscular feedback information, torque estimation and prediction are important to be provided beforehand. Evoked EMG (eEMG) has been found to be highly correlated with FES-induced torque under various muscle conditions, indicating that it can be an useful tool for torque/force prediction. To better construct the relationship between eEMG and stimulated muscular torque, nonlinear-arx-type (NARX-type) model is preferred. This paper presents and exploits a NARX-type recurrent neural network (NARX-RNN) model for identification and prediction of FES-induced muscular dynamics with eEMG. Such NARX-RNN model is with a novel architecture for prediction, with robust prediction performance. To make fast convergence for identification of such NARX-RNN, directly-learning pattern is exploited during the learning phase. Due to difficulty of choosing a proper forgetting factor of Kalman filter for predicting time-variant torque with eEMG, such NARX-RNN may be considered to be a better alternative as torque predictor. Data gathered from two SCI patients is used to evaluate the proposed NARX-RNN model. The NARX-RNN model shows promising estimation and prediction performance only based on eEMG.

We develop and present a portable tool intended for real time estimation of the center of mass (CoM) in human subjects. Using the statically equivalent serial chain (SESC) method we can account for subject specificity after identification of the model's parameters. CoM position estimates are then available from measurements of the subject's limbs orientations. For portability, we make use of widely accessible sensors such as the Kinect and Wii balance board for identification. Use of the Kinect as a measurement device allows us to establish the SESC outside of the laboratory, without many special considerations on the environment. Only Kinect is used for CoM tracking after identification was performed. We present here an overview of the SESC concept and the identification procedure. The aspects involved in the visualization tool are discussed and results are shown in order to verify the performance.

Center of mass (CoM) trajectory is important during standing and walking since it can be used as an index for stability and fall prediction. Unfortunately current methods for CoM estimation require the use of specialized equipment (such as motion capture and force platforms) in controlled environments. This paper aims at applying the statically equivalent serial chain (SESC) method to obtain CoM position using widely available and portable hardware; a Microsoft's Kinect and a Nintendo's Wii balance board. During identification, CoM is approximated by CoP measurements and the virtual chain is created for able-bodied subjects. The result demostrates that the SESC method can be applied outside the laboratory environment using a Kinect. Cross-validation of the identified model was performed to evaluate the accuracy of the method. Results obtained of five subjects are shown and discussed.

In order to understand the human motion control strategies and to restore these functions, or to artificially generate limbs motion it is necessary to have an accurate understanding of the limb dynamics. The inertial parameters can be identify easily, however the joint dynamics is still difficult to model due to the time change with muscle contraction level, fatigue and non-linear dynamics. Using Functional Electrical Stimulation (FES) we propose to identify the joint active dynamics with the pendulum test and to establish a relationship between the level of muscle contraction induced by the stimulation and the visco-elasticity. We measure the data of 2 healthy subjects and propose a model for the knee joint visco-elasticity changes.

A musculoskeletal model that allows to simulate the tremor of wrist joint with three degrees of freedom (DoEs) is developed. The model includes five muscles, extensor carpi radialis brevis, extensor carpi radialis longus, extensor carpi ulnaris, flexor carpi radialis and flexor carpi ulnaris. Simulation of tremor generation based on the 3-DoF model is performed. The tremor disorder can be generated in two directions: flexion-extension and radia-ulnar deviation. Accordingly, experiment is conducted on healthy subjects to verify the feasibility of artificial tremor generation via functional electrical stimulation (FES). Simulation results have shown qualitative agreement with the experimental results.

In patients with complete spinal cord injury, fatigue occurs rapidly and there is no proprioceptive feedback regarding the current muscle condition. Therefore, it is essential to monitor the muscle state and assess the expected muscle response to improve the current FES system toward adaptive force/torque control in the presence of muscle fatigue. Our team implanted neural and epimysial electrodes in a complete paraplegic patient in 1999. We carried out a case study, in the specific case of implanted stimulation, in order to verify the corresponding torque prediction based on stimulus evoked EMG (eEMG) when muscle fatigue is occurring during electrical stimulation. Indeed, in implanted stimulation, the relationship between stimulation parameters and output torques is more stable than external stimulation in which the electrode location strongly affects the quality of the recruitment. Thus, the assumption that changes in the stimulation-torque relationship would be mainly due to muscle fatigue can be made reasonably. The eEMG was proved to be correlated to the generated torque during the continuous stimulation while the frequency of eEMG also decreased during fatigue. The median frequency showed a similar variation trend to the mean absolute value of eEMG. Torque prediction during fatigue-inducing tests was performed based on eEMG in model cross-validation where the model was identified using recruitment test data. The torque prediction, apart from the potentiation period, showed acceptable tracking performances that would enable us to perform adaptive closed-loop control through implanted neural stimulation in the future.

This paper investigates a torque estimation method for muscle fatigue tracking, using stimulus evoked electromyography (eEMG) in the context of a functional electrical stimulation (FES) rehabilitation system. Although FES is able to effectively restore motor function in spinal cord injured (SCI) individuals, its application is inevitably restricted by muscle fatigue. In addition, the sensory feedback indicating fatigue is missing in such patients. Therefore, torque estimation is essential to provide feedback or feedforward signal for adaptive FES control. In this paper, a fatigue-inducing protocol is conducted on five SCI subjects via transcutaneous electrodes under isometric condition, and eEMG signals are collected by surface electrodes. A myoelectrical mechanical muscle model based on the Hammerstein structure with eEMG as model input is employed to capture muscle contraction dynamics. It is demonstrated that the correlation between eEMG and torque is time varying during muscle fatigue. Compared to conventional fixed-parameter models, the adapted-parameter model shows better torque prediction performance in fatiguing muscles. It motivates us to use a Kalman filter with forgetting factor for estimating the time-varying parameters and for tracking muscle fatigue. The assessment with experimental data reveals that the identified eEMG-to-torque model properly predicts fatiguing muscle behavior. Furthermore, the performance of the time-varying parameter estimation is efficient, suggesting that real-time tracking is feasible with a Kalman filter and driven by eEMG sensing in the application of FES.

In this article, we describe an approach to model the electromechanical behavior of the skeletal muscle based on the Huxley formulation. We propose a model that complies with a well established macroscopic behavior of striated muscles where force-length, force-velocity, and Mirsky-Parmley properties are taken into account. These properties are introduced at the microscopic scale and related to a tentative explanation of the phenomena. The method used integrates behavior ranging from the microscopic to the macroscopic scale, and allows the computation of the dynamics of the output force and stiffness controlled by EMG or stimulation parameters. The model can thus be used to simulate and carry out research to develop control strategies using electrical stimulation in the context of rehabilitation. Finally, through animal experiments, we estimated model parameters using a Sigma Point Kalman Filtering technique and dedicated experimental protocols in isometric conditions and demonstrated that the model can accurately simulate individual variations and thus take into account subject dependent behavior.

In this paper, we explore the combined use of inertial sensors and the Kinect for applications on rehabilitation robotics and assistive devices. In view of the deficiencies of each individual system, a new method based on Kalman filtering was developed in order to perform online calibration of sensor errors automatically whenever measurements from Kinect are available. The method was evaluated on experiments involving healthy subjects performing multiple DOF tasks.

Functional electrical stimulation (FES) is effective to restore movement in spinal cord injured (SCI) subjects. Unfortunately, muscle fatigue constrains the application of FES so that output torque feedback is interesting for fatigue compensation. Whereas, inadequacy of torque sensors is another challenge for FES control. Torque estimation is thereby essential in fatigue tracking task for practical FES employment. In this work, the Hammstein cascade with electromyography (EMG) as input is applied to model the myoelectrical mechanical behavior of the stimulated muscle. Kalman filter with forgetting factor is presented to estimate the muscle model and track fatigue. Fatigue inducing protocol was conducted on three SCI subjects through surface electrical stimulation. Assessment in simulation and with experimental data reveals that the muscle model properly fits the muscle behavior well. Moreover, the time-varying parameters tracking performance in simulation is efficient such that real time tracking is feasible with Kalman filter. The fatigue tracking with experimental data further demonstrates that the proposed method is suitable for fatigue tracking as well as adaptive torque prediction at different prediction horizons.

Electrical stimulation (ES) is one of the solutions for drop foot correction. Conventional ES systems deliver predefined stimulation pattern to the affected muscles. However, time-variant muscle response may influence the gait performance as they are difficult to be taken into account in advance. Therefore, closed-loop ES control is important to obtain desired gait in presence of muscle response variation. In this work, a dual predictive control, which consists of two nonlinear generalized predictive controllers, is proposed to track desired torque. The stimulated muscle dynamics are modeled by Hammerstein cascades, with one representing stimulation to activation, the other representing activation to torque. Ankle dorsiflexion torque and ES-evoked EMG of tibialis anterior were recorded experimentally for model identification. The control scheme is validated by following desired torque trajectories with the identified model. The results show that the stimulation pattern obtained from the dual predictive control can produce good torque tracking according to the current muscle condition.

In current biomechanics approach, the assumptions are commonly used in body-segment parameters and muscle strength parameters due to the difficulty in accessing those subject-specific values. Especially in the rehabilitation and sports science where each subject can easily have quite different anthropometry and muscle condition due to disease, age or training history, it would be important to identify those parameters to take benefits correctly from the recent advances in computational musculoskeletal modeling. In this paper, Mass Distribution Identification to improve the joint torque estimation and Muscle Strength Identification to improve the muscle force estimation were performed combined with previously proposed methods in muscle tension optimization. This first result highlights that the reliable muscle force estimation could be extracted after these identifications. The proposed framework toward subject-specific musculoskeletal modeling would contribute to a patient-oriented computational rehabilitation.

Muscle fatigue is an unavoidable problem when electrical stimulation is applied to paralyzed muscles. The detection and compensation of muscle fatigue is essential to avoid movement failure and achieve desired trajectory. This work aims to predict ankle plantar-flexion torque using stimulus evoked EMG (eEMG) during different muscle fatigue states. Five spinal cord injured patients were recruited for this study. An intermittent fatigue protocol was delivered to triceps surae muscle to induce muscle fatigue. A hammerstein model was used to capture the muscle contraction dynamics to represent eEMG-torque relationship. The prediction of ankle torque was based on measured eEMG and past measured or past predicted torque. The latter approach makes it possible to use eEMG as a synthetic force sensor when force measurement is not available in daily use. Some previous researches suggested to use eEMG information directly to detect and predict muscle force during fatigue assuming a fixed relationship between eEMG and generated force. However, we found that the prediction became less precise with the increase of muscle fatigue when fixed parameter model was used. Therefore, we carried out the torque prediction with an adaptive parameters using the latest measurement. The prediction of adapted model was improved with 16.7%-50.8% comparing to the fixed model.

A model-based Functional Electrical Stimulation (FES) would be very helpful for the adaptive movement synthesis of spinal-cord-injured patients. The nonlinearity of the neuromuscular system can be captured through modeling and identification process. However, there are still critical limitations in FES: rapid muscle fatigue and time-varying property. In actual FES, in order to minimize the fatigue, the intermittent stimulation is adopted. In this case, fatigue and recovery occur in sequence. Thus, the time-varying muscle response is really difficult to be predicted for FES force control. In this paper, we propose an identification method to identify unknown internal states and the maximal force parameter which are inside the nonlinear differential equation. Among the internal parameters of muscle model, maximal force Fm should be mainly changed corresponding to the current muscle condition. Muscle fatigue or recovery itself is difficult to be modeled and predicted, however observing the input-output information from the muscle, the adaptive estimation will be achieved to correspond to the varying muscle response effected by a fatigue or unknown metabolic factor of human system. This identification method itself can be expected to be applied for general use in rehabilitation robotics.

Mathematical models of the skeletal muscle can support the development of neuroprostheses to restore functional movements in individuals with motor deficiencies by means of functional electrical stimulation (FES). Since many years, numerous skeletal muscle models have been proposed to express the relationship between muscle activation and generated force. One of them (Makssoud et al.), integrates the Hill model and the physiological one based on Huxley work allowing the muscle activation under FES. We propose in this chapter an improvement of this model by modifying the activation part. These improvements are highlighted through the HuMAnS (Humanoid Motion Analysis and Simulation) toolbox using a 3D biomechanical model of human named Human 36. This chapter describes this toolbox and the software implementation of the model. Then, we present the results of the simulation.

The knowledge and prediction of the behavior of electrically activated muscles are important requisites for the movement restoration by FES in spinal cord injured subjects. The whole parameter's identification of a physiological musculoskeletal model for FES is investigated in this work. The model represents the knee and its associated quadriceps muscle. The identification protocol is noninvasive and based on the in-vivo experiments on paraplegic subjects. The isometric and nonisometric data was obtained by stimulating the quadriceps muscles of 3 paraplegic subjects through surface electrodes. A cross validation has been carried out using nonisometric data set. The normalized RMS errors between the identified model and the measured knee response are presented for each subject.

EMG-to-force estimation for voluntary muscle contraction has many applications in human-machine interaction, motion analysis, and rehabilitation robotics for prosthetic limbs or exoskeletons. EMG-based model can account for a subject&apos;s individual activation patterns to estimate muscle force. For the estimation, so-called Hill-type model has been used in most of the cases. It already has shown its promising performance, but it is still known as a phenomenological model considering only macroscopic physiology. We have already developed the physiological based muscle model for the use of functional electrical stimulation (FES) which can render the myoelectrical property also in microscopic scale. In this paper we discuss EMG-to-force estimation based on this full physiological based muscle model in voluntary contraction. In addition to Hill macroscopic structure, a microscopic physiology originally designed by Huxley is integrated. It has significant meaning to realize the same kind of EMG-to-force estimation with a physiological based model not with a phenomenological model, because it brings the understanding of the internal biophysical dynamics and new insights about neuromuscular activations. Using same EMG data of isometric muscle contraction, the force estimation results are shown by classical approach and new physiological based approach. Its interpretation is also discussed.

EMG-based muscle model has many applications in human-machine interface and rehabilitation robotics. For the muscular force estimation, so-called Hill-type model has been used in most of the cases. It has already shown its promising performance, however it is known as a phenomenological model considering only macroscopic physiology. In this paper, we discuss EMG-force estimation with the full physiology based muscle model in voluntary contraction. In addition to Hill macroscopic representation, a microscopic physiology description as stated by Huxley and Zahalak is integrated. It has significant meaning to realize the same kind of EMG-force estimation with multiscale physiology based model not with a phenomenological Hill model, because it brings the understanding of the internal biophysical dynamics and new insights about neuromuscular activations.

A model-based FES would be very helpful for the adaptive movement synthesis of spinal-cord-injured patients. For the fulfillment, we need a precise skeletal muscle model to predict the force of each muscle. Thus, we have to estimate many unknown parameters in the nonlinear muscle system. The identification process is essential for the realistic force prediction. We previously proposed a mathematical muscle model of skeletal muscle which describes the complex physiological system of skeletal muscle based on the macroscopic Hill-Maxwell and microscopic Huxley concepts. It has an original skeletal muscle model to enable consideration for the muscular masses and the viscous frictions caused by the muscle-tendon complex. In this paper, we present an experimental identification method of biomechanical parameters using Sigma-Point Kalman Filter applied to the nonlinear skeletal muscle model. Result of the identification shows its effective performance. The evaluation is provided by comparing the estimated isometric force with experimental data with the stimulation of the rabbit medial gastrocnemius muscle. This approach has the advantage of fast and robust computation, that can be implemented for online application of FES control.

In laparoscopic surgery, surgeons find particular difficulties related to the operation technique. Due to restricted view, lack of depth information from the monocular endoscope and limited degree of freedom, surgeons find their movements impeded. A support system that provides improved laparoscopic vision would help to overcome the difficulties. If real-time visualization of abdominal structures were feasible, more accurate procedures and improved quantitative evaluations in laparoscopic surgery might be possible. In this study, a laser-scan endoscope system was developed to acquire and visualize the shape and texture of the area of interest instantaneously. The intraoperative geometric information of deformable organ could be applied for robotic safety management via geometric computation of robot position and organ shape. Results of in vivo experiments on a pig liver verified effectiveness of the proposed system. (C) 2006 Elsevier B.V. All rights reserved.

At present, there are representative robot operation systems such as da Vinci and ZEUS which have realized minimally invasive surgery by the use of dexterous manipulators. in the operating room, medical staff must prepare and set up an environment in which the robot has optimal freedom of motion and its functions can be fully demonstrated for every case. The range of motion in which the robot can reach and be maneuvered is restricted by the fixed point of the trocar site. We have developed a preoperative planning system with the function of volume rendering of medical images and automatic positioning by applying an inverse-kinematics computation of surgical robot. The motion of a surgical robot can be simulated in advance with the intuitive interface and kinematics computation program running in the background of the system. If robotic surgery planning with volume rendering of DICOM images is possible, the discussion of a surgical plan can be directly made just after the diagnosis considering the patient-specific structure. This kind of setup platform would be essential for the future introduction of surgical robotics into an operating room. (c) 2006 Elsevier Ireland Ltd. All rights reserved.

In 2001, we started an endoscopic surgery robot for abdominal surgery The concept of the robot is performing various surgical procedures like open surgery in abdominal region without incising body surface. The developed robot has an endoscopic eye and two forceps type manipulators. Using the robot, the surgeon can perform grasping, incising and clipping a soft tissue with cooperative works of the manipulators. We also appended a navigation function to the robot. The function enabled a surgeon to understand a 3D orientation of the robot in an abdominal cavity. After evaluations of the robot using anesthetized pig, we specialized the robot according to various abdominal organs. Currently, we are expanding the concept of the robot and developing a new robotic system for the endovascular surgery. In this paper, we describe an outline of the robot and a future development.

We have been developing the image capturing system using multi-camera known as "Dynamic Spatial Video Camera (DSVC)" system to measure and observe human locomotion quantitatively and freely And a four-dimensional (4D) human model with detailed skeletal structures, joints, muscles, and motor functionality has been built to analyze dynamic inner structure in human motion. The purpose of our research was to estimate skeletal movements from body surface shapes using DSVC and a 4D human model. For this purpose, we constructed a body surface model of a subject and resized the standard 4D human model to match that with geometrical features of the subject's body surface model. Software that integrates the DSVC system and the 4D human model, and allows dynamic skeletal state analysis from body surface movement data was also developed. We applied the developed system in dynamic skeletal state analysis of human in motion and were able to visualize the results of analysis using geometrically resized 4D human model.

We developed a system for measurement of contact pressure at the hip joint surfaces that enables checking of the artificial hip joint condition during surgery. First, we constructed the pressure sensor that forms the artificial joint. We installed eight small pressure sensors to the spherical head component, a part of the ball-socket joint. Next, we developed software for recording and visualizing the detected pressures that were recorded every 1ms. The pressure distribution was displayed with the 3D computer graphics in real-time. The system enabled intuitive recognition of pressure direction 3-dimensions. Next, using the system, we conducted measurements during total hip arthroplasty. Although it requires some improvements in its measurement accuracy, the system allows real-time acquisition of information on the artificial hip joint in real-time. Further improvements of the calibration method should enable more accurate measurements. As a complete system, it will be a useful tool for selecting an appropriate implant that fits a patient's hip joint or for estimating the risk of complications after surgery.

Robotic systems are increasingly being incorporated into general laparoscopic and thoracoscopic surgery to perform procedures such as cholecystectomy and prostatectomy. Robotic assisted surgery allows the surgeon to conduct minimally invasive surgery with increased accuracy and with potential benefits for patients. However, current robotic systems have their limitations. These include the narrow operative field of view, which can make instrument manipulation difficult. Current robotic applications are also tailored to specific surgical procedures. For these reasons, there is an increasing demand on surgeons to master the skills of instrument manipulation and their surgical application within a controlled environment. This study describes the development of a surgical simulator for training and mastering procedures performed with the da Vinci (R) surgical system. The development of a tele-surgery simulator and the construction of a training center are also described, which will enable surgeons to simulate surgery from or in remote places, to collaborate over long distances, and for off-site expert assistance. Copyright (c) 2005 John Wiley & Sons, Ltd.

Laparoscopic surgery including robotic surgery allows the surgeon to be able to conduct minimally invasive surgery. Robotic surgery has been utilized in performing cholecystectomy, laparoscopic prostatectomy and so on. This surgery can provide many benefits for the patient. However, a surgeon is required to master difficult skills for this surgery to compensate for the narrow field of view, limitation of work space, and the lack of depth sensation. To counteract these drawbacks, the purpose of this study is to develop a surgery simulation that provides a method of training and mastering surgical procedures with the da Vinci system. In addition, our system aims to construct a training simulation center that will enable a surgeon to simulate surgery from or in remote places, to collaborate remotely, and to provide guidance from expert surgeons. In this paper, we would like to show the surgery simulation for da Vinci surgery, in particular a cholecystectomy. The integral parts of this system are a soft tissue model which is created by the sphere-filled method enabling real-time deformations based on a patient's data, the operation console for the simulation and the Internet connection. By using this system, a surgeon can perform surgical maneuvers such as pushing, grasping, and detachment in real-time manipulation and we can perform the tele-surgery simulation for training between two Japanese cities. © 2005 CARS and Elsevier B.V.

An endoscopic robot system that we reported at MMVR11 is able to perform various surgical procedures in the stomach by using two manipulators. However, it is difficult for surgeons to recognize the 3D location and the direction of the endoscope&apos;s tip in the abdominal region during robotic surgery. In this research, we have developed a navigation function that enables image-guided surgery by superimposing the patient&apos;s abdominal organ structure onto the endoscopic image. In this paper, we describe the overview of the navigation for the robot system and the result of an animal experiment done while applying the system.

Intra-operative navigation in which the target position is provided to assist an intuitive understanding of the surgical field has been studied and applied in many clinical areas. Position measurement of a surgical field is usually performed with a magnetic sensor, a marker type optical position sensor. For navigation of hard tissue, the measurement of several markers dispersedly located on the surface is enough to detect the position of an object that can be assumed as a rigid body. However, for the navigation of soft tissue such as skin and liver, a sensor that can measure the deformation of the object surface time-sequentially would be essential. We have developed a 3D visualization system for surgical field deformation with geometric pattern projection. In an animal experiment, the registration of preoperative 3D organ model could be done with the time-sequentially updated surface deformation data. In the video image of surgical field, the inner structure model of organ could be superimposed successfully.

Preoperative simulation and planning of surgical robot setups should accompany advanced robotic surgery if their advantages are to be further pursued. Feedback from the planning system will plays an essential role in computer-aided robotic surgery in addition to preoperative detailed geometric information from patient CT/MRI images. Surgical robot setup simulation systems for appropriate trocar site placement have been developed especially for abdominal surgery. The motion of the surgical robot can be simulated and rehearsed with kinematic constraints at the trocar site, and the inverse-kinematics of the robot. Results from simulation using clinical patient data verify the effectiveness of the proposed system.

Laparoscopic surgery including robotic surgery allows the surgeon to be able to conduct minimally invasive surgery. A surgeon is required to master difficult skills for this surgery to compensate for the narrow field of view, limitation of work space, and the lack of depth sensation. To counteract these drawbacks, we have been developing a training simulation system that can allow surgeons to practice and master surgical procedures. In addition, our system aims to distribute a simulation program, to provide a means of collaboration between remote hospitals, and to be able to provide a means for guidance from an expert surgeon. In this paper, we would like to show the surgery simulation for da Vinci (TM) surgery, in particular a cholecystectomy. The integral parts of this system are a soft tissue model which is created by the sphere-filled method enabling real-time deformations based on a patient&apos;s data, force feedback devices known as a PHANToM (TM) and the Internet connection. By using this system a surgeon can perform surgical maneuvers such as pushing, grasping, and detachment in real-time manipulation. Moreover, using the broadband communication, we can perform the tele-surgical simulation for training.

We have been developing a virtual surgery system that can simulate various surgical maneuvers on elastic organs with haptic feedback, and we have constructed an elastic organ model called the sphere-filled model. This paper describes a method for making a deeper incision in the same region and a slanting incision that approaches the resection for volume data by the use of the sphere-filled model. Using this method, we are able to perform not only pushing and pinching but various incisions that approach the resection in this system. © 2004, Elsevier Science B.V. All rights reserved.

Recently, the surgical robot has been applying in clinical situation. However, to achieve this, it is necessary for the surgeon to master and practice the handling of the instrument. This paper describes the development of the tele-training simulator enabling a surgeon to master and practice the techniques for robotic surgery, in particular, the robotic surgery system "da Vinci". In this simulator, the soft tissue model used is a sphere-filled model that is suited for real-time and quantitative deformation. With respect to the haptic device, we used a commercial device known as the "PHANToM" that has gained widespread use. In addition to that, we have tried to construct a training center system enabling the user to simulate surgery from multiple remote access points. As a result, the user can perform a realistic manipulation similar to the da Vinci forceps&apos; operation. In addition, it allows a user to perform surgical maneuvers such as grasping, pushing and ablation. Moreover, using the broadband line via the Internet, we can perform the tele-virtual training simulation. (C) 2004 Published by Elsevier B.V.

We have developed a data fusion system for the robotics surgery system "da Vinci". The data fusion system is composed of an optical 3D location sensor and a digital video processing system. The 3D location sensor is attached to the da Vinci's laparoscope and measures its location and direction. The digital video processing system captures the laparoscope's view and superimposes 3D patient's organ models onto the captured view in real-time. We applied the system to "da Vinci" and examined this system during a cholecystectomy. In this experiment, the surgeon was able to observe the inner conditions of the organs with a stereo view.

Precise measurements of geometry should accompany robotic equipments in operating rooms if their advantages are further pursued. For deforming organs including the liver, intraoperative geometric measurements play an essential role in computer surgery in addition to pre-operative geometric information from CT/MRI. The laser-scan endoscope system acquires and visualizes the shape of the area of interest in a flash of time. Results of in-vivo experiments on the liver of a pig verify the effectiveness of the proposed system. In the next stage, we aim to make a data-fusion in laparoscopy.