Sage Journals: Discover world-class research

Abstract

Wearable robots, especially those composed of soft materials, are increasingly attracting interest due to their comfort, ease of donning and doffing, and their ability to provide assistance across various applications. In wearable robotics, striking a balance between ensuring low impedance for wearer comfort and providing sufficient assistive force is a notable design challenge. In this study, we propose exploiting impedance variation in accordance with the types of muscle contraction in the human body. Particularly in eccentric muscle contraction, the impedance can help reduce the muscular load, since it exerts force in the same direction as the muscles. To utilize the relation, we proposed a linked-layer jamming mechanism, which adjusts its impedance largely in various directions. This mechanism allows not only a broad variable range of impedance in multiple rotation directions but also directional torque design, even when equipped in human multi-degree-of-freedom (DoF) joints. By constructing a wearable robot prototype equipped with the proposed linked-layer jamming mechanisms, the effectiveness of this impedance-based assistance approach was confirmed through experiments. The findings from this study present new possibilities in wearable robot design, showing that suitably amplified impedance can assist human motion, potentially enhancing task efficiency and lowering injury risk. This work thus offers a new perspective for researchers in the field of wearable robots, demonstrating that impedance, often minimized in existing designs, can be utilized beneficially when properly amplified.

I. Introduction

Recently, wearable robots composed of soft materials have attracted great interest due to their ease of donning and doffing, comfort, and ability to provide assistance. They are being researched in a wide variety of fields, ranging from assisting human muscle strength to reducing the load on the muscles of the wearer,^1–4 and even for psychological comfort.⁵ In particular, research on soft wearable robots for assisting the upper limbs is frequently conducted in the rehabilitation and industrial fields.^6–9

While it is important to reduce the impedance of soft wearable robots for comfortable wearing, it is very challenging to ensure sufficient assistive force at the same time. Higher impedance in wearable robots may lead to increased muscular effort, regardless of whether the robot is operating or not.^10–12 Especially in multi-degree-of-freedom (multi-DoF) joints, such as the wrist or shoulder, it is more difficult to reduce the impedance due to the large number of required actuators.^13,14 Therefore, to develop a wearable robot with low impedance and sufficient assistive force, careful consideration of this trade-off relationship is essential in robot design.

By understanding and utilizing the physical nature of impedance, it can be used to assist the motion of the wearer in certain situations. Impedance is a measure of how much a system resists motion, exerting a force in the opposite direction to the motion. In the case of eccentric muscle contraction in the human body, where the muscles lengthen to produce a contractile force, the muscles exert a force in the opposite direction of the movement, such as impedance. Exploiting this commonality, it is possible to reduce the force required on human muscles using a wearable robot to increase the impedance of the body at the appropriate moment. If the assistive effect of “impedance-based assistance” is verified, all existing mechanisms with varying impedance will become more valuable in wearable robots.

The variable impedance mechanism required to implement impedance-based assistance has been explored in several literatures, but it still requires refinement for exploiting in wearable robots that assist the multi-DoF joints of the human body.¹⁵ The variable impedance mechanism using particle jamming is smooth and adaptable to various environments,¹⁶ but it is difficult to apply when the required impedance is different in each direction. It is possible to have unequal impedances in various directions using soft layer jamming¹⁷ or using an origami structure,¹⁸ but these are limited to specific directions of impedance design. A layer jamming mechanism has also been developed to achieve multi-DoF by incorporating joints to layers,¹⁹ but it is not presented in the experimental verification of its performance when embedded in wearable robots. Mechanisms for varying impedance passively have also been studied,²⁰ but they are not suitable for impedance-based assistance, which requires varying impedance at different moments depending on the external force encountered. There are various methods for jamming particles or layers,^21–23 such as tendon-driven or positive pressure; they require a certain volume or thickness for varying the impedance of the mechanisms, which can result in bulky shapes in wearable robots.

In this research, we propose a linked-layer jamming mechanism suitable for multi-DoF soft-wearable robots to implement impedance-based assistance (Fig. 1). The proposed layer jamming mechanism has a wide variable range of impedance in the multi-DoF rotational direction. In addition, a mathematical model of the resistive torque in each direction is constructed to enable directional torque design for human multi-DoF joints. We experimentally verified the effectiveness of impedance-based assistance to the human body by constructing a wearable robot prototype using the proposed linked-layer jamming mechanisms. This result presents new possibilities for wearable robot design by demonstrating that the proper usage of impedance can assist human motion.

FIG. 1.

Developed wearable robot utilizing the linked-layer jamming mechanisms capable of varying impedance in multiple directions.

II. Impedance-Based Assistance

Mechanical impedance can positively affect the motion of a system by fully understanding and utilizing physical principles of the impedance. Impedance is a measure of how much a system resists motion, encompassing properties such as inertia, stiffness, and damping. Generally, researchers attempt to reduce impedance for efficient robot performance. However, impedance can be used to improve the motion of a system by utilizing the principle that impedance generates a force in the opposite direction of motion. For situations that require a force opposing motion, such as changing the direction of motion, maintaining posture, or braking, impedance can be used positively for the system behavior.

Especially in the human body, impedance can be utilized to reduce the muscular force load when a muscle is performing an eccentric contraction. In an eccentric muscle contraction, the muscle produces a contractile force as it lengthens (Fig. 2A). In other words, the muscle is exerting force in the opposite direction of the movement, meaning that the muscle and the impedance are exerting force in the same direction (Fig. 2B). Since muscle and impedance perform the same role during eccentric muscle contraction, increasing the impedance of a human joint reduces the load required on the muscle (Fig. 2C). It can also be verified by the musculoskeletal simulation results in Figure 2D. As the stiffness and damping of the joint increase, the force required for the eccentrically contracting muscles decreases.

FIG. 2.

Principle of impedance-based assistance. (A) Eccentrically contracted muscle exerts the force opposite to the body motion. (B) Impedance exerts the resistive force in opposite direction to the body motion. (C) Thus, impedance can reduce the muscular burden when the muscle contracts eccentrically. (D) Musculoskeletal simulation results from OpenSim²⁴ demonstrate that impedance can be used to assist eccentric muscle contraction.

When assisting eccentric muscle contractions, it is safer to use impedance compared to conventional robotic assistance methods. For assisting eccentric muscle contractions with traditional robots, additional power must be applied in the opposite direction of the joint motion.^8,25 If the robot is overdriven, it can cause an increased burden on the opposing muscle to maintain the desired motion. In severe situations, the direction of motion of the human body can be reversed or the user can be seriously injured. Impedance, in contrast, allows for inherently safe assistance because it eventually halts joint motion even when excessive force is applied.

Eccentric muscle contractions are common in daily tasks, and are particularly prone to muscle damage and injury. Most daily tasks, which involve multi-DoF movements in various joints, often require eccentric contractions in several muscles. For example, in carrying an object, the dominant muscles of the elbow or shoulder eccentrically contract as the object is set down. Despite their frequent use, eccentric muscle contractions are known to cause more muscle fiber damage and slower recovery than concentric contractions.^26,27 If we reduce the load required on the muscle during eccentric contractions, we could potentially improve work efficiency and even prevent injury. As such, eccentric muscle contractions are an essential part of human life, and utilizing impedance to assist them could lead to safer wearable robots.

III. Design of Wearable Robot

We developed a linked layer jamming mechanism suitable for impedance-based assistance on multi-DoF joints (Fig. 3). The bendable layers allow pitch rotation, and the joints between the layers allow yaw rotation with almost frictionless motion. By lowering the internal air pressure, the layers are squeezed together, resulting in increased stiffness and damping, and hence impedance. By analyzing the resistive torque in each rotational direction, it is possible to design the directionally specific impedance.

FIG. 3.

The operational principles of the linked-layer jamming mechanism and the analysis and design for developing it into a wearable robot.

A. Impedance design of linked-layer jamming mechanism

In this section, we analyzed the rotational impedance in the pitch and yaw directions of the linked-layer jamming mechanism to investigate the effect of design parameters.

(1) Pitch rotation

Pitch-directional rotation is achieved by bendable layers. The resistive torque in the pitch rotation of the layers can be expressed as the sum of the bending moment of an elastic beam and friction between the layers.²⁸ When the several layers are stacked, the rotational resistive torque can be expressed as follows: $τ_{pitch, atm} = k θ_{pitch} + c {\dot{θ}}_{pitch}$ (1) $k = \frac{2 E w n t^{3}}{3 L_{layer}^{4}}$ (2) $c = μ Δ P w (n - 1) t$ (3) $= 0 (∵ Δ P = 0)$ (4) $“ τ_{pitch, atm} ”$ represents the resistive torque in pitch direction under atmospheric pressure. $“ k ”$ and $“ c ”$ represent lumped stiffness and damping coefficients in the whole layers along the angle, $“ θ_{pitch,} ”$ and angular velocity, $“ {\dot{θ}}_{pitch,} ”$ in pitch direction. $“ E ”$ is the Young’s modulus of a layer. $“ w, ”$ $“ n, ”$ $“ t, ”$ and $“ L_{layer} ”$ represent the width, number, thickness, and length of the layer. $“ μ ”$ represents a friction coefficient between the layers. $“ Δ P ”$ represents the pressure difference compared with atmospheric pressure (Fig. 4A).

FIG. 4.

Details of design considerations. (A) Illustration of the parameters associated with the linked-layer jamming mechanism. (B) An example of wearing the wearable robot on the elbow and illustrates the changing friction area with movement.

Governing equation in negative pressure can be expressed as follows, which is separated into slip and nonslip conditions.²⁸ $τ_{pitch, vac} = {\begin{matrix} τ_{slip} (τ_{slip} < τ_{non - slip}) \\ τ_{non - slip} (τ_{slip} > τ_{non - slip}) \end{matrix}$ (5) $τ_{slip} = \frac{2 E w n t^{3}}{3 L_{layer}^{4}} θ_{pitch} + c {\dot{θ}}_{pitch}$ (6) $τ_{non - slip} = \frac{2 E w {(n t)}^{3}}{3 L_{layer}^{4}} θ_{pitch}$ (7) $“ τ_{pitch, vac} ”$ represents the resistive torque in pitch direction when negative pressure is engaged. It is separated into $“ τ_{slip} ”$ and $“ τ_{non - slip,} ”$ which represent the resistive torque in slip and nonslip conditions.

In equations, the parameter $“ n ”$ appears effective due to the difference between nonslip conditions and atmospheric conditions. However, although the increase of $“ n ”$ can enlarge the torque in the slip state, it is meaningless due to an increase of the resistive torque in atmospheric conditions. The resistive torque in the nonslip state can be effective for assisting isometric contraction, but not in eccentric contraction accompanying slight motion. Therefore, the parameter $“ n ”$ is partially effective for obtaining wider impedance variation but not in all conditions.

2) Yaw rotation

The linked-layer jamming mechanism allows yaw rotation from the joint between the layers. The resistive torque in the yaw direction is generated by frictional forces on the contact area at the joint. In atmospheric conditions, the resistive torque $“ τ_{yaw, atm} ”$ is negligible although there is slight friction from little contact. $τ_{yaw, atm} ≅ 0$ (8)

In contrast, under negative pressure, the resistive torque in yaw direction $“ τ_{yaw, vac} ”$ is caused by the friction on the layer joint, which can be expressed by the following equation. $τ_{yaw, vac} = n \int \int r F_{friction}$ (9) $= n \int_{0}^{r} \int_{0}^{2 π} r (μ r d θ d r Δ P)$ (10) $= μ n Δ P \int_{0}^{r} 2 π r^{2} d r$ (11) $= \frac{2}{3} μ n π r^{3} Δ P$ (12) $“ F_{friction} ”$ is a frictional force caused in micro area of $“ d θ d r . ” “ r ”$ is the radius of the layer joint in which contact occurs.

Equations (8)–(12) demonstrate that the resistive torque has exhibits a variable range in the yaw direction. It remains close to zero at atmospheric pressure, although it increases proportionally to the third power of the contact area radius, denoted in parameter $“ r, ”$ under negative pressure. In addition, the increase in parameter $“ n ”$ is significantly contributes to the yaw directional impedance, which can also increase the impedance in pitch direction simultaneously. The verification and analysis of the parametric tendency will be covered in the experiment given in Section 4.

In conclusion, impedance in each direction can be adjusted by designing the parameter $“ n ”$ simultaneously and $“ t ”$ and $“ r, ”$ respectively. Due to the low impedance in atmospheric conditions, we can find that the yaw-directional impedance is more suitable for wearable robots. Containing the wearable design in the next subsection, we selected parameters recorded in Table 1.

Table 1.

Parameters of the Linked-Layer Jamming Mechanisms Utilized in the Wearable Robot

Body joint	Motion	Linked-layer jamming mechanism parameter
Body joint	Motion	L [m]	w [mm]	t [mm]	n	$d_{total} [m]$	r [mm]
Shoulder	Adduction / Abduction	0.28	20	1.2	25	0.15	10
Shoulder	Flexion / Extension	0.21	20	1.2	25	0.15	10
Elbow	Flexion / Extension	0.31	30	1.2	20	0.07	15
Wrist	Flexion / Extension	0.12	20	1.2	20	0.03	10
Wrist	Radial / Ulnar Deviation	0.12	20	1.2	10	0.04	10

B. Wearing-oriented wearable robot design

We have developed a soft wearable robot using the linked-layer jamming mechanisms for implementing impedance-based assistance. A main design goal is effectively varying the impedance of the human body using the wearable robot. To achieve it, we have designed the robot involving the potential impedance changes when it is worn by users.

First, the impedance of the wearable robot should be designed to remain unaffected by external pressures from contact or impact. The pressure from external disturbances or the anchoring structure can affect the frictional forces between layers. To prevent it, the area vector of the frictional surface should be perpendicular to the direction of external pressure. We have addressed it by stacking the layers in perpendicular directions to the external disturbance as depicted on the right side of Figure 3.

Second, the wearable robot should be capable of adequately following the movements of joints with varying rotation centers. To address it, robot designers should consider the distance from the joint rotation center to the mechanism (Fig. 4B). ${Δ L}_{\max}_{} = (a_{joint} + \frac{1}{2} w) θ_{joint, \max} < \sum d_{slide}$ (13) $“ {Δ L}_{\max} ”$ represents the maximum required displacement caused by the joint movement, and $“ d_{slide} ”$ is the actuator’s displacement. $“ a_{joint} ”$ is the distance from the center of rotation of the human joint to the skin surface, $“ w ”$ is the width of the layer, and $“ θ_{joint, \max} ”$ is the maximum range of motion of the human joint.

In addition, the variation in the frictional area due to both rotation and translation should also be considered (Fig. 4B). In the case of Equations (3)–(7), the formula for calculating the frictional area applies only when the contact area is circular. To calculate torque for noncircular frictional areas, Equations (9)–(11) can be utilized, tailored to specific shapes. $τ_{yaw, vac, shape} = n \int \int r F_{friction}$ (14) $= n \int \int μ Δ P r^{2} d θ d r$ (15) $= μ n Δ P \int \int \sqrt{x^{2} + y^{2}} d xdy$ (16)

From the equations, rotational resistance in a noncircular friction area can be estimated. For example, when we estimate the resistive torque in the lengthened circular area as depicted in the left box of Figure 4B, the frictional area consists of a square and two semicircular shapes. We can calculate the final torque by adding the additional resistive torque caused by joint rotation-induced translation. $τ_{yaw, total} = τ_{yaw} + τ_{longitudinal}$ (17) $τ_{yaw} = τ_{yaw, square} + 2 τ_{yaw, semicircle}$ (18) $τ_{longitudinal} = n μ A_{friction} Δ P$ (19) $= μ n Δ P (l_{slide} d + π r^{2})$ (20)

Through the above equations, the parameters of the linked-layer jamming mechanism were selected for the prototype (Table 1). The layers were fabricated using three-dimensional printers with polylactic acid filaments. The layers were enclosed by thermoplastic silicon-coated fabric sutured by thermal press (Fig. 5B).

FIG. 5.

Wearable robot using linked-layer jamming mechanisms. (A) is the internal structure of the proposed mechanism, (B) indicates the complete form with a fabric enclosure, and (C) indicates the prototype of the developed wearable robot with the linked-layer jamming mechanisms.

Finally, the multi-DoF assistive wearable robot prototype was constructed using the fabricated linked-layer jamming mechanisms (Fig. 5C). The fabricated jamming mechanisms were equipped on each joint along each direction. In the prototype suit, mechanisms were equipped for assisting 10 DoFs in the human arms.

IV. Experiments

In this section, experiments were conducted to validate the performance and usability of the developed wearable robot. First, we measured and analyzed the rotational impedance of the fabricated linked-layer jamming mechanisms in each direction. In addition, we demonstrated that the impedance of the wearable robot can provide assistance while minimizing interference with other rotations in multi-DoF joints. In single-joint eccentric motion, the auxiliary effect through the impedance of the robot was verified. Furthermore, a lifting task performance experiment demonstrated that the wearable robot can effectively assist in multi-joint and multi-DoF tasks. The experiments were conducted with the approval of the Yonsei University Institutional Review Board (IRB number: 7001988-202305-HR-1917-02).

A. Impedance of the linked-layer jamming mechanism

First, we verified the impedance variation of the proposed linked-layer jamming mechanisms. To measure the sufficient variation in the resistive torque in each rotational direction, we constructed a test bed as depicted in Figure 6A. In the test bed, we estimated the rotational impedance using force-to-displacement data, converted via a pulley with a constant moment arm. The force-to-displacement data were obtained with ASM-1000 force gauge stand (DigiTech Inc.) using DTG-FX300 software. The force gauge pulled a steel wire connected to the pulley with 200 mm/min speeds, up to 80 mm displacement, which causes angular displacement of the pulley at 55 degrees.

FIG. 6.

Resistive torque measurement experiment of the linked-layer jamming mechanism. (A) Depicts the experimental test bed setup. (B) Presents graphs analyzing the variation in pitch rotational resistive torque in relation to the number of layers n and layer thickness t, respectively. (C) Presents graphs analyzing the changes in yaw rotation resistive torque with respect to the number of layers n and the radius r of each layer. At atmospheric pressure, the mechanism maintains low impedance due to low friction. Under vacuum pressure, impedance increases with rising parameters, and distinct impedance profiles emerge in slip and nonslip conditions.

Figure 6B and C present the estimated rotational impedance in the Pitch and Yaw directions, respectively. The results show distinct differences in torque along with the measuring directions and parameter variations. In addition, the results demonstrate clear differences in the inclination of the plots, which indicates stiffness, between the nonslip and slip states.

We analyzed the results in terms of stiffness, using the inclination of torque-to-angle plots at each condition (Table 2). The stiffnesses in a nonslip state, utilized in assisting isometric contraction, dramatically increase in all conditions. In contrast, the stiffness in a slip state, utilized in assisting eccentric contraction, shows a relatively low increase. Even in the slip state, the large increase in stiffness is revealed in lower $“ t ”$ or higher $“ r . ”$ In the case of parameter $“ n, ”$ the difference in stiffness shows opposite tendencies in the Pitch and Yaw directions.

Table 2.

Impedance Analysis Results From Resistive Torque Measurement Experiments of the Linked-Layer Jamming Mechanism

	Pitch rotation						Yaw rotation
	The numberof layers (n)			Layer thickness (t)			The numberof layers (n)			Radius of layer (r)
	10	15	20	0.6 mm	0.8 mm	1.0 mm	10	15	20	5 mm	10 mm	15 mm
$K_{atm, slip}$ $(N \cdot m / \deg)$	0.008	0.0119	0.0216	0.008	0.0408	0.0679	0.0066	0.0055	0.0086	0.0015	0.0049	0.0055
$K_{vac, non - slip}$ $(N \cdot m / \deg)$	0.1176	0.1592	0.2216	0.1042	0.1636	0.2138	0.3339	0.3749	0.3968	0.1576	0.3631	0.4661
$K_{vac, slip}$ $(N \cdot m / \deg)$	0.0190	0.0427	0.0485	0.0177	0.0573	0.0993	0.0341	0.0541	0.0555	0.0018	0.0169	0.0542
$τ_{\max, static}$ $(N \cdot m)$	0.694	0.991	1.356	0.694	1.019	1.311	2.942	3.993	4.695	0.611	2.458	3.946

The opposite tendency is caused by the large increase in Pitch directional impedance even in atmospheric conditions. In the Pitch directions, the stiffness has a large increase along the increase of parameters $“ n ”$ or $“ t ”$ even in atmospheric states. It might be resulted from that the Pitch directional impedance is caused by the bendability of the layers as mentioned in Section 2. In the Yaw direction, in contrast, the mechanism shows quite low stiffness in the atmospheric state, even compared to slip states. Moreover, it maintains low stiffness while parameters are increased.

Based on similar results to the parametric study in the prior section, we can conclude that the selection of layer laminating direction and parameters is adequate. In the Yaw direction, dramatic differences in stiffness were observed between atmospheric and negative pressure conditions. The increase in radius $“ r ”$ of the layers provides substantial design flexibility for Yaw directional impedance during the slip state. Conversely, minimizing layer thickness and the number of layers enables the design of relatively low impedance in the Pitch direction.

B. Multi-DoF impedance experiment

Next, we estimated the joint impedance when the variable impedance mechanism is equipped in the human body. The experiments focused on the wrist joint, which allows multi-DoF motion and is sensitive to impedance changes due to its lightweight.

To estimate the joint impedance, the angular displacement was measured when a constant torque was applied without voluntary contraction. A constant torque was applied to the wrist using a pulley system with a suspended weight of about 2 kg. The angles of the wrist were measured using inertial measurement units (IMU) attached to the test bed. In addition, electromyography (EMG) sensors were attached to the forearm muscles to ensure that the muscles were not activated. As shown in Figure 7B and C, this was performed separately for the wrist flexion and extension (FE) and radial and ulnar deviation (RU) of the wrist. To minimize the effects of the forearm and elbow, a test bed was constructed as depicted in Figure 7A.

FIG. 7.

Multi-degree-of-freedom (DoF) impedance experiment. (A) Illustrates the configuration of the test bed. In (B) and (C), a constant torque is applied to the wrist joint using an additional 2 kg weight. (D) Demonstrates that the wearable robot exhibits higher impedance compared with commercially available wrist guards. (E) Shows that the wearable robot can independently increase the impedance for each DoF.

Figure 7D illustrates the angular displacements caused by impedance when the developed wearable robot is operated, compared to commercially available wrist braces. Compared with commercial braces, joint displacement was decreased by 26.15%p in the FE and 18.25%p in the RU when assisted by the developed robot. When assisted by the robot, the wrist joint exhibited higher impedance than existing commercial protectors, resulting in a stiffer wrist joint against external torque.

The results in Figure 7E demonstrate the impedance variation of the mechanism in the selected direction. The angular displacements in the FE direction remained unchanged even when the impedance in the RU direction was increased, and vice versa. These tendencies are maintained regardless of the impedance in the selected direction.

In conclusion, the developed wearable robot can provide a significant difference in joint impedance when the mechanism is activated, even greater than existing commercial wrist guards. Furthermore, the impedance design of the wearable robot demonstrated direction-selective variation of the impedance in the multi-DoF joints.

C. Single-joint assistance experiment

In this experiment, we validated the impedance-based assistance using the proposed wearable robot. We conducted experiments targeting the shoulder and elbow joints which are frequently injured due to large muscular forces.

To evaluate the assistive effect on a single muscle group, we utilized the metric of maximum isometric force (MIF). MIF refers to the maximum force that muscles can generate while maintaining a specific posture. The decrease in MIF indicates muscle damage and fatigue, making it a comprehensive and practical metric. The decrease in MIF serves as an indirect indicator to confirm muscle damage caused by eccentric exercise.²⁹

We measured and compared the pre- and postexercise MIF in six healthy adult males based on the presence or absence of robotic assistance. In the posture of Figure 8, the MIF was estimated by recording the scale value during the subjects exerting the forces as much as possible. The measurement of MIF was conducted before and after each subject performed high-load exercises, such as dumbbell curls and lateral raises (Fig. 8C, D). The exercises were performed with a weight equivalent to 30% of pre-exercise MIF, and for eccentric exercises, participants were instructed to lift within 1 s and lower for 5 s. The exercise consisted of a total of 3 sets, each comprising 10 repetitions, with sufficient rest intervals. Furthermore, to minimize the influence of muscle fatigue, both comparison groups were given a one-week rest period before conducting the experiments.

FIG. 8.

Single-joint assistance experiment. (A) and (B) depict the measurement of maximum isometric force (MIF). (C) and (D) show the eccentric exercise for the shoulder and elbow joints, respectively. The reduction rate of MIF in (E) the shoulder joint and (F) the elbow joint is presented. The experiment shows the mitigated loss of muscle strength due to assistance provided by the wearable robot’s impedance.

Figure 8E and F shows the average reduction rates of MIF after shoulder and elbow exercises, respectively. For the shoulder, when assisted by the robot in eccentric exercise, the postexercise MIF decreased by an average of 8.81%p. Similarly, for the elbow, the postexercise MIF decreased by an average of 16.11%p. Although there were deviations between subjects, the results of the experiment were found to be significant. This demonstrates the feasibility of impedance-based assistance by the wearable robot during eccentric contractions.

D. Assistance test for field work

An experiment was conducted to validate the assistive effects of the developed wearable robot for multi-joint tasks. Transferring objects is a common multi-joint task frequently required in various industrial settings, often leading to injuries. Furthermore, the appropriate combination of concentric, eccentric, and isometric contractions makes the task suitable for validating the performance of the wearable robot.

The experimental setup and procedure used in the study are illustrated in Figure 9. IMU sensors attached to each joint and muscle, along with EMG sensors, were utilized to measure joint angles and muscle activation. By combining information from these sensors, the eccentric contraction phases of each muscle were identified and used for determining the joint to be augmented in impedance of the wearable robot.

FIG. 9.

Assistance test for field work. (A) and (B) depict the configuration of the test bed. (C) Illustrates the process of the box-carrying task. (D) Compares the inertial measurement units (IMU) data of each joint between the cases with and without the suit, whereas (E) presents the difference in electromyography (EMG) signals throughout the entire task. Despite involving movement in multiple joints, the assistance of the wearable suit resulted in reduced muscle activity in most muscles.

The experimental results are presented in Figure 9D and E. The plots for the joint angles show that the tasks were performed at regulated speeds and ranges. In the measured joint angles and muscle activation data, the main operating muscle groups were revealed, such as the radial deviation of the wrist joint, elbow flexion, and shoulder flexion. In Figure 9E, muscular usages were indirectly inferred with average values of EMG signals in activated muscles. Despite the additional weight from the wearable robot, there was a reduction in muscle usage observed at the assisted wrist, elbow, and shoulder joints.

Through statistical analysis of 7 subjects, it was shown that the muscle use of the shoulder, elbow, and wrist was reduced depending on the presence or absence of assistance (p < 0.05). Compared with the case of nonwearing suits, the assistive effect is relatively lower in the elbow and shoulder (p < 0.1) than wrist (p < 0.05). It might result from relatively low impedance in the wrist joint, which can be easily affected by the external impedance augmentation. Moreover, due to the characteristic of the EMG signal, which is relatively lower in eccentric contraction than in concentric contraction, the reduction in muscular usage might have been measured lower.

Counterintuitively, the assisted joints were observed with slightly more ranges of motion, even though the assistance was derived from the impedance. While this may seem counterintuitive, we observed that the subjects intentionally engaged in larger movements to perform the desired task while being assisted. As a result, the overall muscle usage was lower despite the increased movement that would typically increase muscle usage.

V. Conclusion

In this study, we developed a wearable robot for reducing muscular effort by variable impedance. The proposed robot utilizes a linked-layer jamming mechanism, which can vary its impedance across a wide range in multiple directions. Through the experiments using the robot, we demonstrated that augmenting the joint impedance at eccentric contraction can reduce muscle effort.

The results of this study, which show the possibility of “impedance for assistance,” contribute to conventional variable impedance mechanisms that allow them to be used in reducing muscular usage when embedded in wearable robots. It means that the variable impedance mechanisms can be used not only for active-grounding or active-morphing but also for wearable robots that can assist the human muscles.

The wearable robot presented in this study exploits negative pressure-based layer jamming mechanisms; thus, it achieves lightweight and compactness while providing high assistive force. As the overheads of pneumatic systems such as compressors or air reservoirs can be simplified and minimized utilizing conventional research,³⁰ it has sufficient potential in the development of a light and compact wearable robot.

The limitation of this study is that the control method for negative pressure was binary, operating in an on/off mode. Given the varying torque demands associated with multi-joint tasks in different postures, achieving appropriate control of impedance and utilizing a sensor capable of measuring joint angle and torque³¹ may yield superior outcomes compared with the findings of this study.

Footnotes

Author Disclosure Statement

No interests to disclose.

Funding Information

This work was supported, in part, by the National Research Foundation of Korea (NRF) Grant funded by the Korean government (MSIT) under Grant RS-2023-00208052, in part, by the Ministry of Trade, Industry & Energy (Korea) under the Industrial Technology Innovation Program under Grant 20007058. This research was also supported by the Yonsei University Research Fund of 2023 (2023-22-0075).

Supplementary Material

References

Xiloyannis

, Alicea

, Georgarakis

A-M

, et al. Soft robotic suits: State of the art, core technologies, and open challenges. IEEE Trans Robot, 2022; 38(3):1343–1362.

Bardi

, Gandolla

, Braghin

, et al. Upper limb soft robotic wearable devices: A systematic review. J Neuroeng Rehabil, 2022; 19(1):87.

Thalman

, Artemiadis

. A review of soft wearable robots that provide active assistance: Trends, common actuation methods, fabrication, and applications. Wearable Technol, 2020; 1:e3.

Walsh

. Human-in-the-loop development of soft wearable robots. Nat Rev Mater, 2018; 3(6):78–80.

Yim

, Noguchi

, Tanaka

. A wearable soft robot that can alleviate the pain and fear of the wearer. Sci Rep, 2022; 12(1):17003.

Kang

, Choi

, Lee

, et al. Exo-glove poly II: A polymer-based soft wearable robot for the hand with a tendon-driven actuation system. Soft Robot, 2019; 6(2):214–227.

O’Neill

, et al. Inflatable soft wearable robot for reducing therapist fatigue during upper extremity rehabilitation in severe stroke. IEEE Robotics and Automation Letters, 2020; 5(3):3899–3906.

Zhou

, Hohimer

, Proietti

, et al. Kinematics-based control of an inflatable soft wearable robot for assisting the shoulder of industrial workers. IEEE Robot Autom Lett, 2021; 6(2):2155–2162.

Lee

, Kim

, Park

H-J

, et al. A spine assistive robot with a routed twisted string actuator and a flat-back alleviation mechanism for lumbar-degenerative flat back. IEEE/ASME Trans Mechatron, 2022; 27(6):5185–5196.

10.

Macadam

, Cronin

, Simperingham

. The effects of wearable resistance training on metabolic, kinematic and kinetic variables during walking, running, sprint running and jumping: A systematic review. Sports Med, 2017; 47(5):887–906.

11.

Inacio

, Dipietro

, Visek

, et al. Influence of upper-body external loading on anaerobic exercise performance. The J Strength Conditioning Res, 2011; 25(4):896–902.

12.

Couture

, Simperingham

, Cronin

, et al. Effects of upper and lower body wearable resistance on spatio-temporal and kinetic parameters during running. Sports Biomech, 2020; 19(5):633–651.

13.

Jeong

, Yasir

, Han

, et al. Design of shape memory alloy-based soft wearable robot for assisting wrist motion. Applied Sciences, 2019; 9(19):4025.

14.

Shi

, Dong

, Lin

, et al. Human joint torque estimation based on mechanomyography for upper extremity exosuit. Electronics, 2022; 11(9):1335.

15.

Vanderborght

, Albu-Schaeffer

, Bicchi

, et al. Variable impedance actuators: A review. Robotics Autonomous Systems 61, 2013; 61(12):1601–1614.

16.

Brown

, Rodenberg

, Amend

, et al. Universal robotic gripper based on the jamming of granular material. Proc Natl Acad Sci USA, 2010; 107(44):18809–18814.

17.

Choi

, Corson

, Peiros

, et al. A soft, controllable, high force density linear brake utilizing layer jamming. IEEE Robot Autom Lett, 2018; 3(1):450–457.

18.

Deshpande

, Ho Tse

, Ren

. “Origami-inspired bi-directional soft pneumatic actuator with integrated variable stiffness mechanism.” 2017 18th international conference on advanced robotics (ICAR). IEEE, 2017.

19.

Choi

, Kim

, Lee

, et al. Soft, multi-DoF, variable stiffness mechanism using layer jamming for wearable robots. IEEE Robot Autom Lett, 2019; 4(3):2539–2546.

20.

Yun

S-S

, Kim

, Ahn

, et al. Body-powered variable impedance: An approach to augmenting humans with a passive device by reshaping lifting posture. Sci Robot, 2021; 6(57):eabe1243.

21.

Choi

, Lee

D-Y

, Eo

J-H

, et al. Tendon-driven jamming mechanism for configurable variable stiffness. Soft Robot, 2021; 8(1):109–118.

22.

Liu

, Xia

, Lee

D-Y

, et al. A positive pressure jamming based variable stiffness structure and its application on wearable robots. IEEE Robot Autom Lett, 2021; 6(4):8078–8085.

23.

Wang

, Yang

, Zhou

, et al. Inflatable particle-jammed robotic gripper based on integration of positive pressure and partial filling. Soft Robot, 2022; 9(2):309–323.

24.

Delp

, Anderson

, Arnold

, et al. “OpenSim: Open-source software to create and analyze dynamic simulations of movement. IEEE Trans Biomed Eng, 2007; 54(11):1940–1950.

25.

Lotti

, Xiloyannis

, Durandau

, et al. Adaptive model-based myoelectric control for a soft wearable arm exosuit: A new generation of wearable robot control. IEEE Robot Automat Mag, 2020; 27(1):43–53.

26.

Warren

, Lowe

, Armstrong

. Measurement tools used in the study of eccentric contraction-induced injury. Sports Med, 1999; 27(1):43–59.

27.

Lieber

, Fridén

. Mechanisms of muscle injury after eccentric contraction. J Sci Med Sport, 1999; 2(3):253–265.

28.

Narang

, Vlassak

, Howe

. Mechanically versatile soft machines through laminar jamming. Adv Funct Materials, 2018; 28(17):1707136.

29.

Michaut

, Pousson

, Babault

, et al. Is eccentric exercise-induced torque decrease contraction type dependent? Med Sci Sports Exerc, 2002; 34(6):1003–1008.

30.

Lee

, Lee

, Shin

. An air recirculation system based on bioinspired soft re-air valve for highly efficient pneumatic actuation. Soft Robot, 2021; 8(5):564–576.

31.

Yun

, Kim

, Shin

. A Contact Joint Sensor Using a Force-Sensor Module Embedded in a 3D Curved Surface. IEEE Sensors J, 2022; 22(1):867–878.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

114.40 MB

0.00 MB

Impedance for Assistance: Upper-Limb Assistive Soft Robotic Suit Using Linked-Layer Jamming Mechanisms

Abstract

I. Introduction

II. Impedance-Based Assistance

III. Design of Wearable Robot

A. Impedance design of linked-layer jamming mechanism

(1) Pitch rotation

2) Yaw rotation

B. Wearing-oriented wearable robot design

IV. Experiments

A. Impedance of the linked-layer jamming mechanism

B. Multi-DoF impedance experiment

C. Single-joint assistance experiment

D. Assistance test for field work

V. Conclusion

Footnotes

Author Disclosure Statement

Funding Information

Supplementary Material

References

Supplementary Material