People who bought this also bought...
Beccai, S. Micera, C. Cipriani, J. Carpaneto and M. Roccella, E. Cattin, N.
Wearable Robots | Wiley Online Books
Vitiello, F. Vecchi and M. Tsagarakis, D. Caldwell and S. Moreno, E. Turowska and J. Turouska and J.
Bastos-Filho, M. Sarcinelli-Filho, A. Ferreira, W. Celeste, R. Silva, V. Martins, D. Cavalieri, P. Filgueira and I. Undetected location. NO YES. Home Subjects Nanotechnology General. Wearable Robots: Biomechatronic Exoskeletons. Selected type: Hardcover. Added to Your Shopping Cart. View on Wiley Online Library. This is a dummy description. A wearable robot is a mechatronic system that is designed around the shape and function of the human body, with segments and joints corresponding to those of the person it is externally coupled with.
Teleoperation and power amplification were the first applications, but after recent technological advances the range of application fields has widened. Increasing recognition from the scientific community means that this technology is now employed in telemanipulation, man-amplification, neuromotor control research and rehabilitation, and to assist with impaired human motor control.
About the Author Jose L. Pons has also written Emerging Actuator Technologies: A Micromechatronic Approach a book on the design and control of novel actuators for applications in micro nanosystems. Permissions Request permission to reuse content from this site. Wisse 2. Forner-Cordero 2. Schiele 3. Moll 3. Baydal-Bertomeu 3. Pons 3. Pons 4. Carmena 4. Benini 4. Pons 5. Schiele 5. Vivas 5. Pons 6.
Powered exoskeletons in fiction
Brunetti and J. Pons 7. Vanhala 7. The other exoskeletons can still have a large range of motion of the trunk. However, they do not adopt their kinematics or support with variations in lifting style. Kinematic compatibility in the frontal and transverse planes should also be considered. The Laevo system incorporates rotating elements in the chest pad, which to some degree fulfill the function of a differential transmission. The alignment for the legs is provided through the elasticity of a belt behind the waist.
In comparison, the Robomate exoskeleton features more elaborated misalignment compensation for the hip, consisting of two hinge joints and one ball joint Toxiri et al. The same compensation mechanism is used for the trunk Toxiri et al. The BNDR does not include any mechanism to account for kinematic compatibility Ulrey and Fathallah, other than the hip joints consisting of torsional springs.
For the WMRD exoskeleton it is not clear, if any additional compensation mechanisms are integrated at the hip joint Wehner et al. A rotational joint at the top of the support strut enables axial rotation in the transverse pane. Soft, suit like structures have a relatively long history in back support devices Abdoli-Eramaki et al. Like other exosuits, the forces are transmitted in tension only and no weight bearing structure parallel to the human exists. While the lack of rigid mechanic structures greatly enhances kinematic compatibility, the range of motion can still be a challenge Abdoli-Eramaki et al.
However, Inose et al. To avoid these compression forces, they updated their design to include a flat spring mechanism in the back as a load bearing structure. As the mathematical model from Abdoli-Eramaki et al. Several reasons are reported in the literature as to why back support exoskeletons and exoskeletons in general are not widely adopted in the industry yet: discomfort Barrett and Fathallah, ; Abdoli-Abdoli-Eramaki et al. This paper presents the design and preliminary testing of a passive back support exoskeleton, that allows for a large range of motion of the lumbar spine and the hip and asserts kinematic compatibility with the user See Figure 1B.
In section 2. The concept of the novel back support exoskeleton and its design are explained in section 2. The mechanical implementation of the elastic back support mechanism, the misalignment compensating hip module as well as the passive torque source at the hip are described in section 2. Subsequently, the experimental testing of the components as well as the testing involving human subjects are elaborated in section 3.
The outcomes of the component testing, biomechanical testing and functional testing are presented in section 4 and discussed thereafter in section 5. An ideal back support exoskeleton can reduce the peak and cumulative loading on the back while at the same time still allow for a large range of motion. In order to be effective, a back support exoskeleton needs to connect the torso an the thighs. Due to the complexity of the hip joint and lumbar spine, many degrees of freedom need to be carefully bridged See Figure 2A.
The kinematic compatibility and the fit of the device are tightly linked to the comfort of the user. Both characteristics are discussed in more detail in the following. In the process of designing back support exoskeletons, the degrees of freedom introduced by the lumbar spine are not always taken into account or are merged with the degrees of freedom of the hip Wehner et al. Simple models of the human spine, suggest, that the lumbar spine can be modeled as one additional joint in the sagittal plane see Figure 2A. More complex models, consist of five separate spherical joints to model the lumbar spine Christophy et al.
The angular displacement in the lumbar spine can occur in the sagittal plane by flexing or extending the spine. Lateral bending in the frontal plane and axial rotation in the transverse plane constitute the rest of the possible angular displacements. Peak values for the range of motion of the human hip, spine and trunk can be found in Table 1. Table 1. The hip joint is commonly modeled as a a ball joint Pons, Its degrees of freedom are comprised of flexion and extension in the sagittal plane, abduction and adduction in the frontal plane and internal external rotation in the transverse plane.
The main reason, why the degrees of freedom of the human are important, is the fact, that a misaligned exoskeleton joint can cause some unwanted forces and moments Schiele and van der Helm, Additionally it can lead to unwanted relative movements between the device and the wearer. This problem is well-documented in Schiele and van der Helm and is common in all exoskeleton joints that are not aligned See Figure 2C.
However, an exoskeleton joint can be misaligned without causing harm, as long as it is properly compensated. There are several ways to solve this. One that is used in a number of devices Schiele and van der Helm, ; Schiele, ; Toxiri et al. Anthropomorphic data Tilley, suggests that significant differences exist in hip and waist width See Figure 2B. These differences are in the range of several centimeters. If not properly accounted for, these differences can be a source of discomfort and misalignment.
Bio-mechanical linked segment models Kingma et al. Therefore, designers often decide to compensate for a fraction of the full dynamic torque, which ranges typically between 20 and 30 Nm for purely passive devices Abdoli-Eramaki et al. More complex, optimization based models indicate, that torque requirements for a passive back support exoskeleton could be as low as approximately 20 Nm Millard et al.
However, the toques of an active system for the lumbar spine and hip are significantly higher. In the optimization, the active torques were saturated to 67 Nm. Based on the requirements above there were two main design goals: kinematic compatibility and the support torque in the same order of magnitude as in passive back support exoskeletons of approximately 20—30 Nm, should be transmitted.
Here, first the concept for the spinal part is described, followed by the concept for the hip part. Since it is challenging even for skilled people to correctly align an exoskeleton Schorsch et al. Correct alignment even with single hinge joints, such as the elbow, are a challenge Schiele and van der Helm, Therefore, correct alignment with all five spherical joints of the lumbar spine is arduos. However, since good results can be expected by compensating for misalignment, this approach was chosen here. Additional degrees of freedom can be added in many forms.
One that at the same time allows to store energy, is the use of flexible materials See Figure 3A. A long and slim flexible structure can be bent in two directions and deformed under torsion. However, even to compensate for flexion and extension in the sagittal plane, at least two additional compensating degrees of freedom are required Schiele and van der Helm, One to account for the length change, and one for correct alignment with the back. The same holds for the lateral bending and axial rotation. Figure 3.
Spinal support concept A. The lumbar joint of the exoskeleton consists of a flexible beam, which serves at the same time as a torque source. On the beam of length L acts a force F at the top. This leads to a displacement v x which is a function of the position along the beam x. The flexion extension joint is powered with a Maccepa 2. The pretension P is altered with a worm gear. The main design parameters are an offset B from the center of rotation, here a fixed profile radius R and the length C ; The misalignment compensation of the hip asserts kinematic compatibility with additional three parallel joints and a slider.
The choice of the position of the additional degrees of freedom used to provide compensation is relatively arbitrary, but attention should be payed to singular positions of the used mechanism Schiele, However, the choice of a mechanism can have practical implications Junius et al. Choosing a center of rotation of an exoskeleton joint, to be located relatively far away from the corresponding human joint, will require a big compensatory effort, compared to almost coinciding joints. A flexible structure is used as a compensatory joint and energy storage a the same time. A combination with a linear joint along the flexible structure and an additional spherical joint was found to compensate for a large part of the misalignments in an iterative approach See Figure 4.
Figure 4. Left Passive Spexor prototype. An elastic spinal module in the back is used to provide a large range of motion and to store energy. Several misalignment compensating mechanisms at the hip and the back hip indicated in the figure are implemented. A passive, parallel elastic torque source provides support at the hip; Right Kinematic equivalent representation. The spinal structure is comprised of a ball joint and a linear slider in combination with elastic beams here represented as two hinge joints in series. This structure provides flexibility and compensates for potential misalignment.
The hip structure consists of three parallel hinge joints in combination with a fourth, actuated hinge joint in series with a linear joint. This structure fulfills a dual function, once as a fitting mechanism and once as a misalignment compensating structure. For the flexion and extension of the hip joint, a mechanism with many adjustment possibilities was chosen. Simple models form the literature indicate Toxiri et al.
The goal was to replicate this torque profile with adjustment possibilities See Figure 3B in the torque magnitude Vanderborght et al. Alignment of the rotation axis of a human and an exoskeleton is challenging Junius et al. For the flexion and extension and abduction and abduction, an approach was chosen, where a certain amount of misalignment is accepted and compensated for by the device kinematics See Figure 2E and for more detail Figure S1 , i.
Additionally, a wide range of sizes should be fitted. Based on a simple kinematic model of the human hip and the exoskeleton the workspace in the frontal plane is estimated. Anthropomorphic data from Tilley, and range of motion estimates from Magee, are used to estimate the human workspace See Figure 2F and for more detail Figure S2. The kinematic equations are provided in the Supplementary Material. Especially, when additional compensatory degrees of freedom are introduced.
However, because it is challenging to connect to the thigh in a way, that no rotation around the thigh axis occurs, we decided to use this to our advantage and implement the compensation of the internal and external rotation of the hip this way: the exoskeleton structure stays rigid and the leg rotates inside the not too much tightened cuff. The presence of large muscle groups and compliant tissues of the leg further simplify this approach. Next to the problem to compensate for misalignment, the fitting of the exoskeleton on the hip is essential, especially, if the exoskeleton should not be protruding.
One mechanism, that can be used to accommodate a wide range of different waist to hip widths, is one that consists of three parallel joints in series See Figure 5B. Incidentally, this mechanism can also be used to compensate for a small amount of misalignment.
Figure 5. Elastic spinal module A. This mechanism is on the one hand used to compensate for misalignment and on the other hand to provide a good fit. Torsional springs are mounted on the top two joints to avoid singularities and increase comfort; Misalignment compensation mechanism hip and torque source hip C. The torque, that is generated by the actuator is transmitted through the misalignment compensation mechanism. In the design of a controlled-brake orthosis Durfee and Goldfarb, and later in the Robomate Toxiri et al. The difference to our design is though, that in our case, the three compensatory joints are placed above the flexion-extension joint See Figure 4.
This allows us to use the mechanism additionally for fitting purposes and place the connection to the pelvis very close to the body. Further, this configuration avoids collisions of the exoskeleton with the leg, which could potentially occur for large abduction angles in the controlle-brake orthosis or Robomate exoskeleton. However, this approach also comes with potential disadvantages. By placing the typically actuated and somewhat heavy flexion—extension joint below the three hinge joints for misalignment, the weight of the structure tends to rest against the skin of the wearer.
Exerting a compression force at the hip. This might become a source of discomfort, especially if the actuator is heavy. Therefore, in the presented design, the two top hinge joints are equipped with torsional springs See Figure 5B. Additionally, the two spring loaded joints counteract the tendency of the entire mechanism, to go into a singular configuration along the leg. In order to extend the range of motion even further, especially for lateral bending and axial rotations of the trunk, a spring loaded slider along the leg is added See Figure 4. Potentially, this allows the flexion—extension joint to self align.
In the following, the design considerations of the passive Spexor exoskeleton See Figure 4 are described in more detail. Specifically, the physical interfaces, the spinal module, the misalignment compensation mechanisms, and the torque source at the hip. The pelvis structure, which connects the hip and the spinal part of the exoskeleton, is a custom made carbon fiber frame by Otto Bock HealthCare GmbH, consisting of two separate L-shaped structures that are clamped in the back.
The pelvis structure is therefore adjustable in width. The mechanical loading of the pelvis structure is considerable. In order to prevent too large torsion angles of the structure, additional clamping of the ends of the overlapping structures was added See Figure 5A. The shoulder interface stems from a modified backpack Karrimor Panther An aluminum structure was added to mount the ball joint and to prevent the backpack from introducing unwanted slack into the system.
Therefore, a new concept consisting of multiple continuous carbon fiber beams under bending loads was developed See Figure 5A. Advantages of the continuous beams include, that the overall structure is light weight, compact and comparatively simple. However, one of the main disadvantages is, that excessive loading breaks the beams. The peak stresses occur at the base of the beam, therefore special attention was payed to the fixation at the base. The beams are clamped in a structure made out of 3D printed ABS plastic embedded in an aluminum structure.
At the base of the structure, the ABS plastic is modeled in such a way that the plastic follows the bending curve of the beam. The active length of the beam can be individually adjusted to the size of the user. The beams are cut to a length of mm and are 4 mm in diameter. Three beams in parallel are used to provide a torque in the range of 30 Nm See Figure 5A. In order to dimension the beams, linear and non-linear models were used.
The linear model follows a text book approach and is described in the following. In order to calculate the displacement v x as a function of x along a cantilever beam See Figure 3A the following second order differential equation has to be solved:. The Young's modulus E is a material property, related to stiffness, while the second moment of area I is related to the geometric properties.
For a round cross section beam, like the one used here, the second moment of area amounts to:. For a cantilever beam, where a force F is applied at the top of the beam, the moment along the beam has the following form:. Integrating this equation twice, with the boundary conditions, that a the displacement at the base is zero, i. This equation can be used to explicitly calculate the displacement v x as a function of the force F , the Young's modulus E , the second moment of area I the length L and the position x along the beam.
Qualitatively, the linear and non-linear displacements look similar. The misalignment compensation and fitting mechanism is manufactured out of aluminum See Figure 5B. The distance between two parallel steel axis is 30 mm. Two parallel custom designed torsion springs are mounted on the first two joints. For future testing, the first two joints can be locked with steel pins in discrete positions, additionally, encoder mounts are foreseen on all joints. Inverse pendulum models of the trunk predict a torque profile required around the hip to have a sinusoidal profile.
In order to deliver such a torque profile, a force generated with a linear die compression spring Sodemann ST , with a spring constant of 21 N mm , is routed over a profile disk to generate a non-linear profile with two 2 mm Dyneema cables See Figure 5C. The actuator is designed in such a way, that by changing manually the pretension, the peak output torque can be adjusted in a range of 10—30 Nm per joint, amounting to a total of 20—60 Nm.
The exoskeleton is tested in two ways: the isolated components, and the interaction of the exoskeleton with the user. The user testing is comprised of two parts, biomechanical and functional testing. In order to asses the behavior of the individual components of the exoskeleton the carbon fiber beam and the torque source at the hip were benchmarked with respect to their torque angle characteristic. This allows to estimate the torque provided by the exoskeleton with kinematic data. In order to benchmark the components, they were fixed in a custom made test stand See Figure 6A,B.
The components were equipped with reflective markers and their position was tracked with a camera system Vicon Vero at a frequency of Hz. External forces, recorded with a load cell Futek LSB , N at 1 kHz, were applied manually to excite the components. The results are reported in section 4.quisumanpero.cf
Exoskeleton: A Novel
Figure 6. Component testing: Spinal support A. The beams are tested on the one hand to identify the material properties, such as the Young's modulus E and on the other hand to identify the torque angle characteristic. The carbon fiber beams are fixed on the base. A force, which is recorded with a load cell, is applied at the top.
Five points on the beams are tracked using a motion capture system Vicon Vero. Torque source hip B. In order to validate the design of the torque source and to identify the torque angle characteristic, the hip actuator is subjected to a external force. The force is recorded with a load cell, while at the same time the two markers on the rotating, top part of the actuator are tracked; Loading support verification C.
Subject 1 is wearing the bottom part of the exoskeleton. The exoskeleton is equipped with markers, whose position is recorded with the motion capture system. Subject 2, with known mass m is standing on a force plate, while exciting the spinal part of the exoskeleton with the forces F 1 and F 2. In order to verify the loading support of the single components combined, the torque-angle behavior of the entire exoskeleton was identified.
The exoskeleton was equipped with Optotrak markers, while one user wore the hip part of the exoskeleton See Figure 6C. Instead of connecting the shoulder interface to the user, a second person standing on a force plate would excite the exoskeleton while the wearer flexes out of the way. The experiment was once performed with the flexible beams See Figure 7A and once with the rigid aluminum tube See Figure 7C. The resulting support torque-angle characteristic is reported in section 4.
Figure 7. Different exoskeleton configurations were tested: A Flexible exoskeleton with slider unlocked. B Flexible exoskeleton with slider locked. C Rigid exoskeleton with slider unlocked. D No exoskeleton. Three young, healthy male subjects average age: 30 years, average height: During the experiment, participants were asked to perform a range of motion ROM in all three directions; flexion-extension y , lateral bending x , and rotation z , while holding the knees as straight as possible without locking the knee.
Subjects were asked to put the same amount of effort in each trial when wearing the different exoskeleton configurations See Figure 7. In total, subjects performed four maximal ROM tasks, differing in the configuration of the exoskeleton:. Exo flex-slider: Flexible back structure, with back slider unlocked. Exo flex-no slider: Flexible back structure, with back slider locked. Exo rigid: Rigid back structure, slider unlocked.
No Exo: Not wearing any exoskeleton. In addition, marker clusters were attached to relevant parts of the exoskeleton; exo hip joint 1 , exo pelvis frame 2 , exo base spinal structure 3 , and exo top spinal structure 4. For all modeled human body segments foot, shank, thigh, pelvis, abdomen, thorax, head, upper arm, forearm, and hand , anatomical coordinate systems were calculated based on anatomical landmarks that were related to the corresponding marker clusters using a probe with six markers Cappozzo et al.
Figure 8. Angle definitions for the human and the exoskeleton A. Knee angle a , hip angle b , lumbar angle c , trunk angle d , pelvis inclination e , exo hip angle f , exo lumbar angle g , exo trunk angle h , exo pelvis inclination i , Measurement setup B. Subject equipped with the exoskeleton and markers to track the position and orientation of different body parts red. The position and orientation of the lower leg A , upper leg B, not visible in this photograph , the pelvis C , the trunk D the upper arm E , and the lower arm F are tracked.
Additionally, the exoskeleton is equipped with marker clusters to track the exoskeleton yellow hip angle 1 the pelvis carbon fiber frame at the hip 2 the pelvis carbon fiber frame at the base of the spinal exoskeleton structure 3 top of the spinal exoskeleton structure 4. Various single markers were used to track individual parts of the exoskeleton, but are not further used in the analysis and therefore not further elaborated.
The stick figures show a visualization of the tracking results. For the sake of clarity, the body markers red are not highlighted in this figures. In the middle the subject is standing upright. The pink lines represent body parts of the subject, while the turquoise color represents the exoskeleton structure. On the right, the subject is performing a stoop.
From the rotation matrices of the segments, joint angles were calculated using Euler decomposition YXZ. The angles that were used in the analysis are defined in Figure 8A. From the angle time-series See Figure 10 , peak angles were subtracted and compared between conditions. Three men with no low-back pain history participated in the functional testing of the protoytpe.
Participants performed a series of functional tasks, including 1 tasks to assess the supportive function of the device, 2 tasks to assess the extent of restricted hip flexion, and 3 tasks to asses the ROM. A detailed description of the tasks can be found in a previous study of Baltrusch et al. For convenience the tasks are also listed in Figure 12B. The functional tasks were performed once with an exoskeleton and without any exoskeleton. The sequence of the tasks and the starting conditions were randomized to prevent order and habituation effects. Functional performance was assessed by using questionnaires after each task, asking for perceived task difficulty and discomfort of the device.
At the end of the session participants had to fill in a user's impression questionnaire regarding their experience with the exoskeleton during the test session. The subjective outcomes measures were all assessed by using a visual-analog scale VAS.