F.C.T. van der Helm

A Series Elastic- and Bowden-Cable-Based Actuation System for Use as Torque Actuator in Exoskeleton-Type Robots

Within the context of impedance controlled exoskeletons, common actuators have important drawbacks. Either the actuators are heavy, have a complex structure or are poor torque sources, due to gearing or heavy nonlinearity. Considering our application, an impedance controlled gait rehabilitation robot for treadmill-training, we designed an actuation system that might avoid these drawbacks. It combines a lightweight joint and a simple structure with adequate torque source quality. It consists of a servomotor, a flexible Bowden cable transmission, and a force feedback loop based on a series elastic element. A basic model was developed that is shown to describe the basic dynamics of the actuator well enough for design purpose. Further measurements show that performance is sufficient for use in a gait rehabilitation robot. The demanded force tracking bandwidths were met: 11 Hz bandwidth for the full force range (demanded 4 Hz) and 20 Hz bandwidth for smaller force range (demanded 12 Hz). The mechanical output impedance of the actuator could be reduced to hardly perceptible level. Maxima of about 0.7 Nm peaks for 4 Hz imposed motions appeared, corresponding to less than 2.5% of the maximal force output. These peaks were caused by the stick friction in the Bowden cables. Spring stiffness variation showed that both a too stiff and a too compliant spring can worsen performance. A stiff spring reduces the maximum allowable controller gain. The relatively low control gain then causes a larger effect of stick in the force output, resulting in a less smooth output in general. Low spring stiffness, on the other side, decreases the performance of the system, because saturation will occur sooner.

Kinematic Design to Improve Ergonomics in Human Machine Interaction

This paper introduces a novel kinematic design paradigm for ergonomic human machine interaction. Goals for optimal design are formulated generically and applied to the mechanical design of an upper-arm exoskeleton. A nine degree-of-freedom (DOF) model of the human arm kinematics is presented and used to develop, test, and optimize the kinematic structure of an human arm interfacing exoskeleton. The resulting device can interact with an unprecedented portion of the natural limb workspace, including motions in the shoulder-girdle, shoulder, elbow, and the wrist. The exoskeleton does not require alignment to the human joint axes, yet is able to actuate each DOF of our redundant limb unambiguously and without reaching into singularities. The device is comfortable to wear and does not create residual forces if misalignments exist. Implemented in a rehabilitation robot, the design features of the exoskeleton could enable longer lasting training sessions, training of fully natural tasks such as activities of daily living and shorter dress-on and dress-off times. Results from inter-subject experiments with a prototype are presented, that verify usability over the entire workspace of the human arm, including shoulder and shoulder girdle

Self-Aligning Exoskeleton Axes Through Decoupling of Joint Rotations and Translations

To automatically align exoskeleton axes to human anatomical axes, we propose to decouple the joint rotations from the joint translations. Decoupling can reduce setup times and painful misalignment forces, at the cost of increased mechanical complexity and movement inertia. The decoupling approach was applied to the Dampace and Limpact exoskeletons.

Design of a series elastic- and Bowden cable-based actuation system for use as torque-actuator in exoskeleton-type training

Common actuators have important drawbacks for use in an exoskeleton type of rehabilitation (training) robot. Either the actuators are heavy, complex or poor torque sources. A new actuation system is proposed and tested that combines a lightweight joint and a simple structure with adequate torque source quality. It consists of a servomotor, a flexible Bowden cable transmission, and a force feedback loop based on a series elastic element. Measurements show that performance is sufficient for use in a gait rehabilitation robot.

Multiple-step strategies to recover from stumbling perturbations.

This study has analysed the recovery from an induced stumble whilst walking on a treadmill. Four stumbling conditions were tested; at early swing with short and long durations and at mid and late swing with short duration. The experiment set-up, including the possibility of being stumbled, did not alter the normal gait patterns and the recovery strategies depended on the perturbation conditions. For the early swing perturbation, delayed lowering and elevating strategies were performed using the perturbed leg. A lowering strategy was seen for mid and late swing perturbations. An elevating strategy consisted of an elevation of the swing limb while a lowering one consisted of bringing the foot quickly to the ground. There were two groups of reactions to the experimental perturbation of gait. In the first, there was an effort to complete the disturbed step as normally as possible, so the following steps were less constrained to maintain treadmill speed. In the second group of reactions, the perturbed step was aborted and the recovery effort transferred to the contralateral limb. In many cases, several steps were needed to regain normal gait pattern. The study of recovery reactions from gait perturbations should include at least three steps after the perturbed one.

Dampace: dynamic force-coordination trainer for the upper extremities

According to reviews, training with upper-extremities rehabilitation robotics is at least as good as regular stroke rehabilitation, probably because the robotics increase the training intensity for the patients. As an alternative to the functional approach mimicking activities of daily living, targeted force-coordination training may also have its benefits. Our passive exoskeleton, the Dampace, has controlled braking on the three rotational axes of the shoulder and one of the elbow. It is designed to combine functional training of activities of daily living with force-coordination training. The Dampace exoskeleton can assist in identifying causes behind the movement disorders of stroke patients, tackle these causes with isolated force-coordination training, possibly simultaneously over multiple joints, and then integrate the isolated training back into a functional, task-specific training protocol. Not needing to align the Dampace axes to the human shoulder and elbow axes overcome some of the difficulties traditionally associated with exoskeletons. Although it adds more complexity, the reduction of setup times to a few minutes and the absence of static reaction forces in the human joints, are major advantages and have been well received by therapists and physicians. Controlled braking instead of actively assisting actuators, has the advantage of inherent safety and always actively participating patients, at the cost of not being able to assist movements or create all virtual environments.

Design of a Rotational Hydroelastic Actuator for a Powered Exoskeleton for Upper Limb Rehabilitation

The goal of this study was to validate the suitability of a novel rotational hydroelastic actuator (rHEA) for use in our new rehabilitation exoskeleton for the upper limbs, the Limpact. The rHEA consists of a rotational hydraulic actuator and a custom-designed symmetric torsion spring in a series-elastic configuration. For rehabilitation therapy and impairment quantification, both compliant impedance control and stiff admittance control modes are possible. In the validation experiments, the torque bandwidth of the rHEA was limited to 18 Hz for a desired 20 N·m reference signal (multisine, constant spectrum) due the transport delays in the long flexible tubes between the valve and cylinder. These transport delays also required changes to existing theoretical models to better fit the models on the measured frequency response functions. The (theoretical) measurable torque resolution was better than 0.01 N ·m and the (validated) delivered torque resolution below 1 N· m. After the validation experiments, further iterative improvements resulted in a spring design capable of a maximum output torque of 50 N·m with an intrinsic stiffness of 150 N· m/rad and a slightly higher bandwidth. With the design locked, the maximum measurable isometric torque is 100 N ·m. In conclusion, the rHEA is suitable for upper limb rehabilitation therapy as it matches the desired performance.

Quantifying Proprioceptive Reflexes During Position Control of the Human Arm

This study aimed to analyse the dynamic properties of the muscle spindle feedback system of shoulder muscles during a posture task. External continuous force disturbances were applied at the hand while subjects had to minimize their hand displacements. The results were analysed using two frequency response functions (FRFs) from which the model parameters were derived, being 1) the mechanical admittance and 2) the reflexive impedance. These FRFs were analysed by a neuromusculoskeletal model that implicitly separates the reflexive feedback properties (position, velocity and acceleration feedback gains) from intrinsic muscle visco-elasticity. The results show substantial changes in estimated reflex gains under conditions of variable bandwidth of the applied force disturbance or variable degrees of external damping. Position and velocity feedback gains were relatively larger when the force disturbance contained only low frequencies. With increasing damping of the environment, acceleration feedback gain decreased, velocity feedback gain remained almost constant and position feedback gain increased. It is concluded that under the aforementioned circumstances, the reflex system increases its gains to maximize the mechanical resistance to external force disturbances while preserving sufficient stability.

Sensory weighting of force and position feedback in human motor control tasks.

In daily life humans integrate force and position feedback from mechanoreceptors, proprioception, and vision. With handling relatively soft, elastic objects, force and position are related and can be integrated to improve the accuracy of an estimate of either one. Sensory weighting between different sensory systems (e.g., vision and proprioception) has been extensively studied. This study investigated whether similar weighting can be found within the proprioceptive sensory system, more specifically between the modalities force and position. We hypothesized that sensory weighting is governed by object stiffness: position feedback is weighted heavier on soft objects (large deflections), while force feedback is weighted heavier on stiff objects (small deflections). Subjects were instructed to blindly reproduce either position or force while holding a one degree of freedom haptic manipulator that simulated a linear spring with one of four predetermined stiffnesses. In catch trials the spring was covertly replaced by a nonlinear spring. The difference in force (ΔF) and position (ΔX) between the regular and the catch trials revealed the sensory weighting between force and position feedback. A maximum likelihood estimation model predicted that: (1) task instruction did not affect the outcome measures, and (2) force feedback is weighted heavier with increasing object stiffness as was hypothesized. Both effects were found experimentally, and the subjects sensory weighting closely resembled the optimal model predictions. To conclude, this study successfully demonstrated sensory weighting within the proprioceptive system.

Energy efficient and robust rhythmic limb movement by central pattern generators

Humans show great energy efficiency and robustness in rhythmic tasks, such as walking and arm swinging. In this study a mathematical model of rhythmic limb movement is presented, which shows that tight local coupling of Central Pattern Generators (CPGs) to limbs could explain part of this behavior. Afferent feedback to flexor and extensor centers of the CPG is crucial in providing energy efficiency by means of resonance tuning. Feedback of positional information provides resonance tuning above the endogenous frequency of the CPG. Integral feedback provides resonance tuning at and below the endogenous frequency. Feedback of velocity information is necessary to compensate for the time delay in the loop, coupling limb to CPG; without velocity feedback bi-stability occurs and resonance tuning is not possible at high movement frequencies. The concepts of energy efficient and robust control of rhythmic limb movements are also applicable to robotics. It is the first CPG model, which provides resonance tuning at natural limb frequencies above and below its endogenous frequency.