SCASPER: Scallop-Shaped Pneumatic Actuator

• The paper introduces SCASPER as a high-force, bi-directional actuator tailored for shoulder actuation in upper-limb rehabilitation.
• SCASPER employs a scallop-inspired, stacked airbag design that amplifies angular motion and maximizes output torque.
• The actuator features rapid, simplified fabrication and is validated through quasi-static modeling and experimental tests on a dummy arm.

Searching arXiv for the specified paper to ground the article in the source record. The Scallop-Shaped Pneumatic Robot (SCASPER) is a soft pneumatic actuator introduced for upper-limb exoskeleton and assistive rehabilitation, specifically for shoulder actuation in a joint-specific design paradigm that pairs SCASPER with the Lobster-Inspired Silicone Pneumatic Robot (LISPER) for the elbow. In "Novel bio-inspired soft actuators for upper-limb exoskeletons: design, fabrication and feasibility study" (Zhang et al., 1 Sep 2025), SCASPER is characterized by high output force/torque and simplified fabrication processes. The actuator is proposed in response to limitations reported for many soft rehabilitation robots, including slow response times, restricted range of motion, low output force, and limited articulation of how bellow-structured actuator designs quantitatively contribute to capability. Within that framework, SCASPER functions as a high-force, bi-directional, geometrically tunable actuator intended for shoulder-level loads, where much larger assistive torque is required than at the elbow.

1. Functional role in upper-limb rehabilitation

SCASPER is assigned to the shoulder, while LISPER is assigned to the elbow (Zhang et al., 1 Sep 2025). The separation is not incidental: the paper explicitly motivates SCASPER by the observation that the shoulder typically requires much larger assistive torque than the elbow. Accordingly, SCASPER is intended to provide dynamic gravitational compensation and assistive shoulder motion in upper-limb rehabilitation, with explicit suitability for shoulder flexion/extension assistance, gravity compensation for a heavy human arm, and integration into a 2-DOF upper-limb assistive system with LISPER.

The design is framed as a response to recurring shortcomings in prior soft shoulder actuators. The paper argues that previous soft shoulder actuators are often bulky, limited in output force, slow to deflate, difficult to fabricate, or too complex for practical wearable rehabilitation devices. SCASPER is proposed to address those issues through a configuration that is high-force, bi-directional, geometrically tunable, faster and simpler to fabricate, and suitable for shoulder-level loads.

A central implication of this design assignment is that the paper advances a joint-specific actuator paradigm rather than a uniform architecture for the entire limb. This suggests that force, range of motion, and fabrication constraints are treated as joint-dependent design variables rather than globally optimized objectives.

2. Structural design, materials, and scallop-inspired geometry

SCASPER is an inextensible airbag pneumatic actuator (Zhang et al., 1 Sep 2025). Its construction uses polyethylene heat-shrinking tubes as the main airbag material, silicone rubber strips connecting the distal edges of the airbags, ABS plates to seal the ends and hold the inlet, PU tubing for air supply, pipe positioning rings for stacking and alignment, and bolts and nuts for fastening via drilled holes.

The choice of polyethylene heat-shrink tubing is a defining material decision. The paper states that it is low-cost, easy to obtain, wear-resistant, compliant enough for deformation, intrinsically suitable for airtight bag formation, and does not require laser cutters, sewing machines, or heat-sealing machines. This establishes SCASPER as fabrication-simple in contrast to many textile-based pneumatic actuators.

Geometrically, SCASPER consists of six stacked airbags arranged in a rotational pattern. Each airbag is a rectangular pouch with two external corners trimmed off, distal and proximal side holes for fixing, and silicone rubber strips at the distal end. The scallop inspiration derives from how a scallop shell opens: the structure starts compact, and internal pressure causes the fans or segments to open outward, producing a large angular extension.

The trimmed corners are functionally significant. The paper states that they move the effective contact points closer to the rotation axis, which increases the achievable extension angle $\theta$. This geometry also helps create a more useful moment arm for generating torque. Although SCASPER is not presented as a classic silicone bellow actuator like LISPER, its opening is treated as a structured unfolding or bending process in which bag geometry determines how pressure turns into angular displacement, the stacked pouch architecture creates a large lever effect, and the silicone strips provide a restoring tendency and improve repeatability and linearity.

The paper summarizes the consequences of this scallop-shaped or bellow-structured configuration as four linked effects: amplifying angular motion, increasing output torque, supporting bi-directional return, and reducing fabrication complexity. A plausible implication is that the biological analogy is being used less as literal biomimesis than as a geometric heuristic for pressure-driven angular deployment.

3. Fabrication process and manufacturability

SCASPER can be fabricated in about 45 minutes excluding airtightness testing (Zhang et al., 1 Sep 2025). The reported fabrication sequence is: cut polyethylene heat-shrink tubing into six equal-length pieces; trim the two distal corners of each airbag; drill 5-hole and 6-hole patterns for fasteners; add inlet connections using PU tubing; seal distal and proximal ends with flat ABS plates; stack the six airbags using positioning rings; merge the six inlets into a single pneumatic line; and connect silicone rubber strips to the distal end of each airbag.

The manufacturability claim is integral to the actuator’s identity. The paper emphasizes that SCASPER requires only scissors or cutting, taping, screwing, and 3D-printed support parts, while avoiding laser cutting, sewing, and heat sealing. It also allows individual bag airtightness checks before full assembly and uses commonly available materials. These features are presented as major design distinctions rather than incidental conveniences.

For wearable rehabilitation devices, fabrication simplicity has direct methodological significance. It reduces barriers to iterative prototyping, leakage inspection, and modular replacement. The paper’s description also suggests that SCASPER’s stacked architecture supports assembly-level debugging in a way that monolithic soft actuators may not.

4. Analytical model and geometry-dependent actuation

The paper presents a quasi-static analytical model relating pressure, bending or extension angle, and torque (Zhang et al., 1 Sep 2025). For the unloaded actuator, the extension angle is fit from FEM results as a polynomial in pressure:

$\Theta(P)=0.0145 P^2+3.0507 P-1.1438$

where $\Theta$ is the total extension angle of SCASPER and $P$ is inflation pressure.

The torque model is introduced from a work and volume relation:

$\Delta V P=\Delta W$

where $\Delta V$ is change in airbag volume, $P$ is pressure, and $\Delta W$ is work done on the actuator by the environment or load. Under the assumption that the volume change is proportional to angle change,

$\Delta V \propto \Delta \Theta$

the derivation proceeds as

$$\Delta V P = F \cdot \Delta \Theta \cdot R = \Delta \Theta \cdot \tau$$

where $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$0 is external force, $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$1 is the lever arm or moment arm, and $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$2 is torque. This leads to

$\Theta(P)=0.0145 P^2+3.0507 P-1.1438$3

and, after substitution of the geometric approximation,

$\Theta(P)=0.0145 P^2+3.0507 P-1.1438$4

where $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$5 is a geometric length term, $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$6 is an actuator length parameter, $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$7 is the moment arm from the contact point to the rotation center, and $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$8 is pressure.

The paper explicitly interprets this expression as implying that torque scales with pressure, effective length, and moment arm, and notes that in this idealized model torque is not directly dependent on $\Theta(P)=0.0145 P^2+3.0507 P-1.1438$9. That conclusion is presented as a useful consequence of the work-geometry derivation.

A further correction is introduced for pipe bending resistance via an Euler–Bernoulli bending model:

$\Theta$0

where $\Theta$1 is the resisting moment from the pipes, $\Theta$2 is Young’s modulus of the PU pipe, $\Theta$3 is the second moment of area, and $\Theta$4 are radii of curvature of the bent pipe segments. The moment of inertia is defined as

$\Theta$5

where $\Theta$6 is inner diameter and $\Theta$7 is outer diameter. The net torque is then

$\Theta$8

which the paper describes as the torque available after subtracting the pipe’s elastic resistance.

The broader design message is that SCASPER performance can be tuned by geometry. The paper identifies contact point location, moment arm, airbag length, number of bags, strip elasticity, corner trimming, and bag spacing as influential variables. Its main qualitative effects are stated directly: closer contact points to the rotation axis lead to larger effective angular extension; a more favorable lever arm leads to larger torque; a stacked airbag arrangement yields larger total deformation and output force; the silicone strip pattern provides better retraction, lower hysteresis, and adjustable linearity; and the inextensible membrane yields higher force and torque than silicone soft actuators.

5. Experimental characterization

SCASPER was evaluated in a mechatronic test system alongside LISPER (Zhang et al., 1 Sep 2025). The hardware included a compressed air source stabilized at 200 kPa, a proportional electronic pressure regulator, an NI myRIO-1900 as master controller, a NUCLEO-F091RC board, JY62 IMUs at both ends of each actuator, and a high-accuracy digital force gauge for force and torque measurement. The experimental categories were pressure versus angle, pressure versus force or torque under constrained angles, bandwidth or sinusoidal tracking, dummy-arm feasibility test, and position control and gravity compensation.

The pressure-angle test for SCASPER was conducted from 10 kPa to 90 kPa in 10 kPa increments. From the configuration table, the reported specifications are as follows:

Parameter	Reported value
Range of motion	122.5°
Maximum pressure	150 kPa
Weight	183 g
Dimensions	122 × 91 × 132 mm
Maximum output	49.5 N / 5.5 Nm

The paper states that the experimental and simulated angles match reasonably well at lower pressures but diverge at higher pressures due to deformation of the attached air tube, mismatch between idealized and real geometry, and stress concentration effects. It further reports that the discrepancy in extension angle became large after about 30 kPa, reaching a maximum angular difference of 43.63° at 70 kPa.

Force and torque measurements were performed at fixed angles of 0°, 45°, and 60°. The key quantitative result is a maximum measured torque of about 5.45 Nm. This is emphasized because SCASPER is intended for shoulder actuation, where large torques are needed.

Bandwidth-related characterization used sinusoidal set-angle tracking at 1 Hz, 0.5 Hz, and 0.25 Hz. The paper reports the following results:

Condition	Frequency	Reported result
Without rubber pattern	1 Hz	range of motion 89.53°, mean time error 0.42 s, max angular error 32.54°
Without rubber pattern	0.5 Hz	range 90.12°, mean time error 0.31 s, max error 29.74°
Without rubber pattern	0.25 Hz	range 98.25°, mean time error 0.34 s, max error 21.54°
With rubber pattern	1 Hz	range 87.33°, mean time error 0.40 s, max error 29.22°
With rubber pattern	0.5 Hz	range 82.32°, mean time error 0.38 s, max error 25.45°
With rubber pattern	0.25 Hz	range 81.55°, mean time error 0.30 s, max error 17.56°

The paper interprets these data as indicating usable dynamic response, moderate latency, somewhat large peak angular errors at higher speed, and improved behavior with the rubber strip pattern, especially in reducing error. The remaining error is attributed to solenoid valve response limits, chamber design inefficiency, and quasi-static modeling assumptions.

6. Dummy-arm feasibility study and control performance

The principal system-level demonstration uses a 2-DOF 3D-printed human dummy arm, with SCASPER mounted at the shoulder and LISPER mounted at the elbow (Zhang et al., 1 Sep 2025). Two test classes were performed: a maximum bending angle test and a position control or gravity compensation test.

In the maximum bending test, the actuators were inflated in open loop. The paper notes LISPER at 100 kPa and SCASPER at 90 kPa or similar test conditions, and reports that LISPER rotates 47.4° under 80 kPa while SCASPER rotates 54.4° under 100 kPa. The end-effector displacement was 10.4 cm horizontally and 32.7 cm vertically. The paper treats this as evidence that the actuator pair can move a limb segment through a meaningful range.

The position control test specified desired angle ranges of −10° to 30° for the elbow and 16° to 60° for the shoulder. The reported maximum deviation was about 8.5° at the elbow and about 10.2° at the shoulder. The trajectories were said to be tracked closely overall, with errors mainly near turning points due to valve response limits.

The gravity compensation test used an inverse quasi-static model to maintain several arm postures against gravity. This matters because the intended use case is assistance in holding or moving the arm rather than generating isolated actuator motion. A plausible implication is that SCASPER’s value is best evaluated at the exoskeleton level, where static and quasi-static support tasks are central.

7. Comparison with LISPER, reported advantages, and limitations

The paper explicitly contrasts SCASPER with LISPER (Zhang et al., 1 Sep 2025). LISPER, inspired by a lobster body or fins, is optimized for elbow or wrist-like joints and emphasizes high linearity, higher bandwidth, precise bending, fast response, and moderate force output; it uses silicone rubber bellows, braces, meshes, and a C-shaped constraint. SCASPER, inspired by scallop opening motion, is optimized for shoulder actuation and emphasizes much higher force or torque, simpler fabrication, bi-directional motion, and tunable linearity through silicone strips; it uses inextensible airbag construction.

The stated tradeoff is clear: LISPER provides better responsiveness and linearity, whereas SCASPER provides much better force and torque for the shoulder and is simpler to make. This comparative framing is essential to the paper’s broader claim that upper-limb soft exoskeletons benefit from joint-specific actuator design rather than one generic actuator for every joint.

The advantages claimed for SCASPER are high force and torque output, fabrication simplicity, good suitability for shoulder actuation, adjustable linearity, and better practicality for wearable rehabilitation. These claims are specified in concrete terms: measured torque around 5.5 Nm; no sewing, laser cutting, or heat sealing; around 45 minutes fabrication time; shoulder-specific design for larger moment demand; tunability via silicone rubber strip pattern; and practical features such as low-cost construction, lightweight form, ease of leakage inspection, and a modular stacked design.

The paper also states several limitations. These include relatively large angular errors during fast sinusoidal tracking, actuator lag due to solenoid valves, simplified analytical assumptions, discrepancy between model and experiment at higher pressures, the need for improved interface design to prevent sliding on the human body, and the need for future work for full human-in-the-loop validation. These limitations constrain interpretation of the feasibility results: the dummy-arm study supports the concept, but not yet full clinical or human-wearable validation.

Taken together, SCASPER is presented as the paper’s high-torque, fabrication-friendly shoulder actuator within a joint-specific soft exoskeleton architecture. The central contribution is not only the actuator itself, but the argument that shoulder and elbow assistance should be realized through different soft robotic morphologies, each tuned to the distinct mechanical demands of its target joint (Zhang et al., 1 Sep 2025).