Kinematic coordinations capture learning during human-exoskeleton interaction.

Human-exoskeleton interactions have the potential to bring about changes in human behavior for physical rehabilitation or skill augmentation. Despite significant advances in the design and control of these robots, their application to human training remains limited. The key obstacles to the design of such training paradigms are the prediction of human-exoskeleton interaction effects and the selection of interaction control to affect human behavior. In this article, we present a method to elucidate behavioral changes in the human-exoskeleton system and identify expert behaviors correlated with a task goal. Specifically, we observe the joint coordinations of the robot, also referred to as kinematic coordination behaviors, that emerge from human-exoskeleton interaction during learning. We demonstrate the use of kinematic coordination behaviors with two task domains through a set of three human-subject studies. We find that participants (1) learn novel tasks within the exoskeleton environment, (2) demonstrate similarity of coordination during successful movements within participants, (3) learn to leverage these coordination behaviors to maximize success within participants, and (4) tend to converge to similar coordinations for a given task strategy across participants. At a high level, we identify task-specific joint coordinations that are used by different experts for a given task goal. These coordinations can be quantified by observing experts and the similarity to these coordinations can act as a measure of learning over the course of training for novices. The observed expert coordinations may further be used in the design of adaptive robot interactions aimed at teaching a participant the expert behaviors.

Impact of Gravity Compensation on Upper Extremity Movements in Harmony Exoskeleton.

Robots have been used to offset the limb weight through gravity compensation in upper body rehabilitation to delineate the effects of loss of strength and loss of dexterity, which are two common forms of post-stroke impairments. In this paper, we explored the impact of this anti-gravity support on the quality of movement during reaching and coordinated arm movements in a pilot study with two participants with chronic stroke. The subjects donned the Harmony exoskeleton which supported proper shoulder coordination in addition to providing gravity compensation. Participants had previously taken part in seven one-hour sessions with the Harmony exoskeleton, performing six sets of passive-stretching and active exercises. Pre- and post-training sessions included assessments of two separate tasks, planar reaching and a set of six coordinated arm movements, in two conditions, outside of and supported by the exoskeleton. The movements were recorded using an optical motion capture system and analyzed using spectral arc length (SPARC) and straight line deviation to quantify movement smoothness and quality. We observed that gravity compensation resulted in an increased smoothness for the subject with high level of impairment whereas compensation resulted in a reduction in smoothness for the subject with low level of impairment in the reaching task. Both participants demonstrated better coordination of the shoulder-arm joint with gravity compensation. This result motivates further studies into the role of gravity compensation during coordinated movement training and rehabilitation interventions.

Human-Like Endtip Stiffness Modulation Inspires Dexterous Manipulation With Robotic Hands.

We present a novel method for biomechanically inspired mechanical and control design by quantifying stable manipulation regions in 3D space for tendon-driven systems. Using this method, we present an analysis of the stiffness properties for a human-like index finger and thumb. Although some studies have previously evaluated biomechanical stiffness for grasping and manipulation, no prior works have evaluated the effect of anatomical stiffness parameters throughout the reachable workspace of the index finger or thumb. The passive stiffness model of biomechanically accurate tendon-driven human-like fingers enables analysis of conservatively passive stable regions. The passive stiffness model of the index finger shows that the greatest stiffness ellipsoid volume is aligned to efficiently oppose the anatomical thumb. The thumb model reveals that the greatest stiffness aligns with abduction/adduction near the index finger and shifts to align with the flexion axes for more efficient opposition of the ring or little fingers. Based on these models, biomechanically inspired stiffness controllers that efficiently utilize the underlying stiffness properties while maximizing task criteria can be developed. Trajectory tracking tasks are experimentally tested on the index finger to show the effect of stiffness and stability boundaries on performance.

The Effects of EMG-Based Classification and Robot Control Method on User's Neuromuscular Effort during Real-Time Assistive Hand Exoskeleton Operation.

EMG-based intention recognition and assistive device control are often developed separately, which can lead to the unintended consequence of requiring excessive muscular effort and fatigue during operation. In this paper, we address two important aspects of the performance of an integrated EMG-based assistive system. Firstly, we investigate the effects of muscular effort on EMG-based classification and robot control. Secondly, we propose a robot control solution that reduces muscular effort required in assisted dynamic daily tasks compared to the state-of-the-art control methods.

Assessment of Upper-Extremity Joint Angles Using Harmony Exoskeleton.

The biomechanical complexity of the human shoulder, while critical for functionality, poses a challenge for objective assessment during sensorimotor rehabilitation. With built-in sensing capabilities, robotic exoskeletons have the potential to serve as tools for both intervention and assessment. The bilateral upper-extremity Harmony exoskeleton is capable of full shoulder articulation, forearm flexion-extension, and wrist pronation-supination motions. The goal of this paper is to characterize Harmony's anatomical joint angle tracking accuracy towards its use as an assessment tool. We evaluated the agreement between anatomical joint angles estimated from the robot's sensor data and optical motion capture markers attached to the human user. In 9 healthy participants we examined 6 upper-extremity joint angles, including shoulder girdle angles, across 4 different motions, varying active/passive motion of the user and physical constraint of the trunk. We observed mostly good to excellent levels of agreement between measurement systems with for shoulder and distal joints, magnitudes of average discrepancies varying from 0.43° to 16.03° and width of LoAs ranging between 9.44° and 41.91°. Slopes were between 1.03 and 1.43 with r > 0.9 for shoulder and distal joints. Regression analysis suggested that discrepancies observed between measured robot and human motions were primarily due to relative motion associated with soft tissue deformation. The results suggest that the Harmony exoskeleton is capable of providing accurate measurements of arm and shoulder joint kinematics. These findings may lead to robot-assisted assessment and intervention of one of the most complex joint structures in the human body.

A Modular Design for Distributed Measurement of Human-Robot Interaction Forces in Wearable Devices.

Measurement of interaction forces distributed across the attachment interface in wearable devices is critical for understanding ergonomic physical human-robot interaction (pHRI). The main challenges in sensorization of pHRI interfaces are (i) capturing the fine nature of force transmission from compliant human tissue onto rigid surfaces in the wearable device and (ii) utilizing a low-cost and easily implementable design that can be adapted for a variety of human interfaces. This paper addresses both challenges and presents a modular sensing panel that uses force-sensing resistors (FSRs) combined with robust electrical and mechanical integration principles that result in a reliable solution for distributed load measurement. The design is demonstrated through an upper-arm cuff, which uses 24 sensing panels, in conjunction with the Harmony exoskeleton. Validation of the design with controlled loading of the sensorized cuff proves the viability of FSRs in an interface sensing solution. Preliminary experiments with a human subject highlight the value of distributed interface force measurement in recognizing the factors that influence ergonomic pHRI and elucidating their effects. The modular design and low cost of the sensing panel lend themselves to extension of this approach for studying ergonomics in a variety of wearable applications with the goal of achieving safe, comfortable, and effective human-robot interaction.

Using a System-Based Monitoring Paradigm to Assess Fatigue during Submaximal Static Exercise of the Elbow Extensor Muscles.

Current methods for evaluating fatigue separately assess intramuscular changes in individual muscles from corresponding alterations in movement output. The purpose of this study is to investigate if a system-based monitoring paradigm, which quantifies how the dynamic relationship between the activity from multiple muscles and force changes over time, produces a viable metric for assessing fatigue. Improvements made to the paradigm to facilitate online fatigue assessment are also discussed. Eight participants performed a static elbow extension task until exhaustion, while surface electromyography (sEMG) and force data were recorded. A dynamic time-series model mapped instantaneous features extracted from sEMG signals of multiple synergistic muscles to extension force. A metric, called the Freshness Similarity Index (FSI), was calculated using statistical analysis of modeling errors to reveal time-dependent changes in the dynamic model indicative of performance degradation. The FSI revealed strong, significant within-individual associations with two well-accepted measures of fatigue, maximum voluntary contraction (MVC) force (rrm=-0.86) and ratings of perceived exertion (RPE) (rrm=0.87), substantiating the viability of a system-based monitoring paradigm for assessing fatigue. These findings provide the first direct and quantitative link between a system-based performance degradation metric and traditional measures of fatigue.

Analytical Design of a Pneumatic Elastomer Robot with Deterministically Adjusted Stiffness.

This paper presents a novel Pneumatic Elastomer Robot (PER), called Deterministically Adjusted Stiffness-Pneumatic Elasotmer Robot (DAS-PER), that can concurrently display preprogrammed elongation and bending behaviors. Our design methodology integrates a comprehensive analytical modeling and additive manufacturing-based fabrication to (i) address current ad-hoc and arduous PERs' fabrication limitations, and (ii) enable deterministic stiffness and deformation behavior tuning based on the desired application. To thoroughly evaluate the efficacy of the presented modeling and fabrication approaches, based on the developed model, we first designed and fabricated two DAS-PERs with different bending and elongation stiffnesses. Next, we performed experimental studies to thoroughly evaluate and compare the expected and obtained deformation behaviors. Results demonstrated the efficacy of the fabrication procedure and model fidelity for successful tunability of DAS-PERs solely based on adjusting two internal structure diameter parameters.

Exploring the Capabilities of Harmony for Upper-Limb Stroke Therapy.

Harmony is a bimanual upper-limb exoskeleton designed for post-stroke rehabilitation. It moves the subject's shoulders and arms through their entire ranges of motion while maintaining natural coordination, is capable of force/torque control of each joint, and is equipped with sensors to measure motions and interaction forces. With these capabilities Harmony has the potential to assess motor function and create individualized therapy regimens. As a first step, five stroke survivors underwent rehabilitation sessions practicing multijoint movements with the device. Each participant performed a total of 1130 motions over seven hours of therapy with no adverse effects reported by participants or the attending therapist, supporting the suitability of Harmony for use in a clinical setting. Donning and doffing time averaged 3.5 minutes and decreased with therapist experience. Reported levels of stress, anxiety, and pain indicate that the Harmony safely assisted in the completion of the trained movements and has great potential to motivate and engage patients. We developed a novel methodology for assessing coordination capability and results from the study indicate that Harmony can enable therapists to identify neuromuscular weakness and maladaptive coordination patterns and develop targeted interventions to address these aspects of upper-limb function. The results suggest Harmony's feasibility and show promising improvements, motivating future study to gain statistical support.

Assessing Wrist Movement With Robotic Devices.

Robotic devices have been proposed to meet the rising need for high intensity, long duration, and goal-oriented therapy required to regain motor function after neurological injury. Complementing this application, exoskeletons can augment traditional clinical assessments through precise, repeatable measurements of joint angles and movement quality. These measures assume that exoskeletons are making accurate joint measurements with a negligible effect on movement. For the coupled and coordinated joints of the wrist and hand, the validity of these two assumptions cannot be established by characterizing the device in isolation. To examine these assumptions, we conducted three user-in-the-loop experiments with able-bodied participants. First, we compared robotic measurements to an accepted modality to determine the validity of joint- and trajectory-level measurements. Then, we compared those movements to movements without the device to investigate the effects of device dynamic properties on wrist movement characteristics. Last, we investigated the effect of the device on coordination with a redundant, coordinated pointing task with the wrist and hand. For all experiments, smoothness characteristics were preserved in the robotic kinematic measurement and only marginally impacted by robot dynamics, validating the exoskeletons for use as assessment devices. Stemming from these results, we propose design guidelines for exoskeletal assessment devices.

Design and characterization of the OpenWrist: A robotic wrist exoskeleton for coordinated hand-wrist rehabilitation.

Robotic devices have been clinically verified for use in long duration and high intensity rehabilitation needed for motor recovery after neurological injury. Targeted and coordinated hand and wrist therapy, often overlooked in rehabilitation robotics, is required to regain the ability to perform activities of daily living. To this end, a new coupled hand-wrist exoskeleton has been designed. This paper details the design of the wrist module and several human-related considerations made to maximize its potential as a coordinated hand-wrist device. The serial wrist mechanism has been engineered to facilitate donning and doffing for impaired subjects and to insure compatibility with the hand module in virtual and assisted grasping tasks. Several other practical requirements have also been addressed, including device ergonomics, clinician-friendliness, and ambidextrous reconfigurability. The wrist module's capabilities as a rehabilitation device are quantified experimentally in terms of functional workspace and dynamic properties. Specifically, the device possesses favorable performance in terms of range of motion, torque output, friction, and closed-loop position bandwidth when compared with existing devices. The presented wrist module's performance and operational considerations support its use in a wide range of future clinical investigations.

Methodologies for determining minimal grasping requirements and sensor locations for sEMG-based assistive hand orthosis for SCI patients.

In this paper, we address two of the most important challenges in development and control of assistive hand orthosis. First, supported by experimental results, we present a method to determine an optimal set of grasping poses, essential for grasping daily objects. Second, we present a method for determining the minimal number of surface EMG sensors and their locations to carry out EMG-based intention recognition and to control the assistive device by differentiating between the hand poses.

Estimating anatomical wrist joint motion with a robotic exoskeleton.

Robotic exoskeletons can provide the high intensity, long duration targeted therapeutic interventions required for regaining motor function lost as a result of neurological injury. Quantitative measurements by exoskeletons have been proposed as measures of rehabilitative outcomes. Exoskeletons, in contrast to end effector designs, have the potential to provide a direct mapping between human and robot joints. This mapping rests on the assumption that anatomical axes and robot axes are aligned well, and that movement within the exoskeleton is negligible. These assumptions hold well for simple one degree-of-freedom joints, but may not be valid for multi-articular joints with unique musculoskeletal properties such as the wrist. This paper presents an experiment comparing robot joint kinematic measurements from an exoskeleton to anatomical joint angles measured with a motion capture system. Joint-space position measurements and task-space smoothness metrics were compared between the two measurement modalities. The experimental results quantify the error between joint-level position measurements, and show that exoskeleton kinematic measurements preserve smoothness characteristics found in anatomical measures of wrist movements.

Variable Thumb Moment Arm Modeling and Thumb-Tip Force Production of a Human-Like Robotic Hand.

The anatomically correct testbed (ACT) hand mechanically simulates the musculoskeletal structure of the fingers and thumb of the human hand. In this work, we analyze the muscle moment arms (MAs) and thumb-tip force vectors in the ACT thumb in order to compare the ACT thumb's mechanical structure to the human thumb. Motion data are used to determine joint angle-dependent MA models, and thumb-tip three-dimensional (3D) force vectors are experimentally analyzed when forces are applied to individual muscles. Results are presented for both a nominal ACT thumb model designed to match human MAs and an adjusted model that more closely replicates human-like thumb-tip forces. The results confirm that the ACT thumb is capable of faithfully representing human musculoskeletal structure and muscle functionality. Using the ACT hand as a physical simulation platform allows us to gain a better understanding of the underlying biomechanical and neuromuscular properties of the human hand to ultimately inform the design and control of robotic and prosthetic hands.

A Simulation Framework for Virtual Prototyping of Robotic Exoskeletons.

A number of robotic exoskeletons are being developed to provide rehabilitation interventions for those with movement disabilities. We present a systematic framework that allows for virtual prototyping (i.e., design, control, and experimentation (i.e. design, control, and experimentation) of robotic exoskeletons. The framework merges computational musculoskeletal analyses with simulation-based design techniques which allows for exoskeleton design and control algorithm optimization. We introduce biomechanical, morphological, and controller measures to optimize the exoskeleton performance. A major advantage of the framework is that it provides a platform for carrying out hypothesis-driven virtual experiments to quantify device performance and rehabilitation progress. To illustrate the efficacy of the framework, we present a case study wherein the design and analysis of an index finger exoskeleton is carried out using the proposed framework.

Statistical method for prediction of gait kinematics with Gaussian process regression.

We propose a novel methodology for predicting human gait pattern kinematics based on a statistical and stochastic approach using a method called Gaussian process regression (GPR). We selected 14 body parameters that significantly affect the gait pattern and 14 joint motions that represent gait kinematics. The body parameter and gait kinematics data were recorded from 113 subjects by anthropometric measurements and a motion capture system. We generated a regression model with GPR for gait pattern prediction and built a stochastic function mapping from body parameters to gait kinematics based on the database and GPR, and validated the model with a cross validation method. The function can not only produce trajectories for the joint motions associated with gait kinematics, but can also estimate the associated uncertainties. Our approach results in a novel, low-cost and subject-specific method for predicting gait kinematics with only the subject's body parameters as the necessary input, and also enables a comprehensive understanding of the correlation and uncertainty between body parameters and gait kinematics.

A novel framework for virtual prototyping of rehabilitation exoskeletons.

Human-worn rehabilitation exoskeletons have the potential to make therapeutic exercises increasingly accessible to disabled individuals while reducing the cost and labor involved in rehabilitation therapy. In this work, we propose a novel human-model-in-the-loop framework for virtual prototyping (design, control and experimentation) of rehabilitation exoskeletons by merging computational musculoskeletal analysis with simulation-based design techniques. The framework allows to iteratively optimize design and control algorithm of an exoskeleton using simulation. We introduce biomechanical, morphological, and controller measures to quantify the performance of the device for optimization study. Furthermore, the framework allows one to carry out virtual experiments for testing specific "what-if" scenarios to quantify device performance and recovery progress. To illustrate the application of the framework, we present a case study wherein the design and analysis of an index-finger exoskeleton is carried out using the proposed framework.

Muscle-tendon units provide limited contributions to the passive stiffness of the index finger metacarpophalangeal joint.

The passive stiffness at the MCP joint is a result of the elasticity of muscle-tendon units (MTUs) and capsule ligament complex (CLC), however, the relative contributions of these two components are unknown. We hypothesize that the MTUs provide the majority of the contributions to the joint stiffness by generating resistive forces when the MCP joint is flexed or extended. We used the work done by passive moments as a measure for the determination of the contributions to the joint stiffness. We conducted experiments with ten human subjects and collected joint angle and finger tip force data. The total passive moment and joint angle data were fitted with a double exponential model, and the passive moments due to the MTUs were determined by developing subject-specific models of the passive force-length change relationships. Our results show that for all the subjects, the work done by the passive moments from the MTUs is less than 50% of the total work done, and the CLC provides dominant contributions to the joint stiffness throughout the flexion-extension range of the joint angle. Therefore, the hypothesis that the MTUs provide the majority of the contributions to the MCP joint stiffness is not supported. We also determined that the majority of the MTUs passive moment was generated by the extrinsic MTUs and the contributions of the intrinsic MTUs was negligible.

Contributions of intrinsic visco-elastic torques during planar index finger and wrist movements.

Human hand movements have been studied for many decades, yet the role of hand biomechanics in achieving dexterity has not been fully understood. In this paper, we investigate the contributions of the intrinsic passive viscoelastic component in the hand during the coordinated wrist and hand movements. We compare the contributions of stiffness, damping, and dynamics torques under two types of joint phase movements at two speeds. The analysis of the data collected from subject studies demonstrated that the passive visco-elastic component is dominant over dynamic coupling terms. Although the exact contributions of the three torques vary under different speeds and phasic movements, the stiffness torque was the highest (at least 47%) followed by the damping torque, while the dynamics torque was the lowest (less than 11%) in all movement scenarios. Comparisons with studies involving coordinated arm movements illustrate that dominant torques in arm and hand movements are different suggesting that neural control strategies might be distinct as well.

Acquiring variable moment arms for index finger using a robotic testbed.

Human level of dexterity has not been duplicated in a robotic form to date. Dexterity is achieved in part due to the biomechanical structure of the human body and in part due to the neural control of movement. We have developed an anatomically correct testbed (ACT) hand to investigate the importance and behavioral consequences of anatomical features and neural control strategies of the human hand. One of the critical aspects of understanding dexterity is the analysis of the relationships between the hand muscle movements and joint movements, defined by the moment arms of the muscles. It is known that the moment arms for the hand muscles are configuration-dependent and vary substantially with change in posture. This paper presents a methodology for determining continuous variations in the moment arms with respect to multiple joints moving simultaneously. To determine variations in the moment arms of the ACT hand index finger muscles, we employed a nonparametric regression method called Gaussian processes (GPs). GPs give a functional mapping between the joint angles and muscle excursions, and the gradients of these mappings are the muscle moment arms. We compared the moment arm relationships of the ACT hand with those determined from the available cadaver data. We present the implications of the determination of variable moment arms toward understanding of the biomechanical properties of the human hand and for the neuromuscular control for the ACT hand index finger movements.