Self-Wearing Adaptive Garment (SWAG)

• Self-Wearing Adaptive Garment (SWAG) is a soft robotic dressing system that uses pneumatic subvines to unfurl a pre-inverted fabric sheath for autonomous garment application.
• The design balances force capacity with compliance by optimizing subvine count, channel geometry, and flexible fabric to safely negotiate human joints.
• Experimental results demonstrate rapid dressing (approximately 15 seconds) across diverse configurations with 100% success, emphasizing its efficiency and adaptability.

The Self-Wearing Adaptive Garment (SWAG) is a soft robotic dressing system that uses an unfurling and growth mechanism to facilitate autonomous dressing for individuals with limited mobility. In contrast to robotic dressing approaches that predominantly rely on rigid robotic manipulators, SWAG embeds pneumatically actuated subvines inside a thin fabric sheath that is pre-inverted inside itself. During deployment, the sheath unfurls tip-first over an arm, leg, or garment opening, thereby eliminating skin–garment friction, reducing posture constraints, and enabling a safer and more efficient dressing process across multiple garment configurations (Kim et al., 9 Jul 2025).

1. Conceptual basis within robotic dressing assistance

Robotic dressing assistance has been proposed as a means to improve quality of life for individuals with limited mobility. Existing solutions predominantly rely on rigid robotic manipulators, and the reported limitations motivating SWAG are challenges in handling deformable garments, ensuring safe physical interaction with the human body, excessive operation times, complex control strategies, and constrained user postures (Kim et al., 9 Jul 2025).

SWAG is presented as an alternative architecture in which the garment itself becomes the active dressing mechanism. Rather than transporting fabric over the body by external manipulation, the system conforms to the human body through an unfurling-based deployment method. The central distinction is that new garment material appears at the distal end during deployment instead of sliding over the skin or underlying clothing. This operational principle is the basis for the claim that SWAG eliminates skin–garment friction and can support posture-agnostic dressing.

The name “Self-Wearing Adaptive Garment” reflects two linked properties. “Self-wearing” denotes autonomous garment application driven by embedded soft robotic actuation. “Adaptive” denotes the capacity of the same subvine mechanism to follow body contours and to be pre-patterned into different garment topologies, including sleeves, jackets, and pants.

2. Unfurling mechanism and force transmission

The unfurling mechanism is based on a thin fabric sheath, formed as a cylinder and pre-inverted inside itself, with circumferential channels ultrasonically welded into the sheath. Within each channel sits an inner membrane, termed a subvine. Upon pneumatic inflation, each subvine everts from its base and pushes against the tip of its channel. Because the outer sheath is free to invert forward, it unfurls tip-first over the arm or leg with almost no sliding against skin or underlying clothing (Kim et al., 9 Jul 2025).

The mechanics of a single subvine–sheath interaction are modeled as a frictional, non-linear pulley. Torque equilibrium about the pulley axis is written as

$F_v\,a \;=\; F_u\,(a+b)\;+\;f\,,$

where $F_v$ is the eversion force generated by the subvine, $F_u$ is the unfurling force on the sheath, $a$ and $b$ are lever arms to the subvine entry and sheath tip, and $f$ is a residual term lumping internal friction and material deformation. The subvine force is pressure-driven,

$F_v = P\,A\,,$

with $P$ the pneumatic pressure and $A$ the subvine cross-sectional area. Substitution yields the per-subvine unfurling force,

$F_u = \frac{a}{a+b}\,\bigl(P\,A - f\bigr)\,.$

For $F_v$0 identical subvines supporting a garment mass $F_v$1, the global force balance against gravity becomes

$F_v$2

These relations define the mapping from internal pressure to lifting capability. The details further restate this relationship as

$F_v$3

In practice, $F_v$4 is measured empirically as approximately $F_v$5–$F_v$6 of $F_v$7 and varies with bend angle. The significance of the model is that pressure, subvine count, and geometry enter explicitly into the available unfurling force, permitting design trade-offs between actuation capacity and compliance.

3. Compliance, morphology, and fabrication

Compliance is a primary design variable because the device must negotiate human joints without stalling or exerting excessive contact forces. As the number of subvines increases, their radial offset $F_v$8 from the bending axis increases the composite area moment of inertia,

$F_v$9

where $F_u$0 is the subvine diameter (Kim et al., 9 Jul 2025).

The reported design guideline is to use $F_u$1 subvines for dressing arms and legs. With two or fewer subvines, SWAG remains highly compliant about a known major bending axis; with three or more subvines, the system attains full circumferential symmetry but at the cost of higher stiffness. The two-subvine design is therefore described as balancing force capacity, as given by the global force model, against the compliance required to traverse joints.

Both sheath and subvines are made from 30 denier TPU-coated ripstop nylon. The fabric braid orientation is set at $F_u$2 to the cylinder axis to minimize bending stiffness while maintaining good axial tensile strength. The sheath is fabricated by cutting rectangular fabric, folding it into a cylinder, and welding three longitudinal channels. The subvines are fabricated from a narrower strip of the same fabric, folded and ultrasonically welded to form two independent, preloaded cylindrical tubes.

The implemented layout employs two subvines of diameter $F_u$3 inside a sheath of nominal diameter $F_u$4–$F_u$5. Welding is ultrasonic, with three parallel seams for sheath channels and one for each subvine tube. Typical TPU thickness is $F_u$6–$F_u$7, and operating pressures are up to $F_u$8, chosen to stay well under material tensile limits.

A compact base station contains an acrylic pressure chamber with air inlet from an external compressor, a motor-driven spool using a Dynamixel XL330 for subvine retraction, and a reel to collect the subvine during doffing. The morphology of the actuator and the base station together define SWAG as a garment-integrated soft robotic system rather than an external manipulator with an attached textile end effector.

4. Actuation and control architecture

Actuation is pneumatic. Each subvine is inflated from the base station, and inflation pressure is the primary control input. Retraction is handled by reversing the motor spool, while the channel walls prevent buckling of the retracted subvine (Kim et al., 9 Jul 2025).

The present prototype uses open-loop pressure control and has no onboard sensors. The pressure setpoint is increased until the distal sheath tip reaches the desired coverage point, determined either by time or manual observation. For research experiments, a load cell can measure unfurling force, and a pressure sensor can record peak pressure during joint negotiation.

The proposed feedback extensions remain prospective. Adding a pressure-feedback loop would enable automatic detection of peak pressure, interpreted as indicating tip clearance past a joint. A simple state machine could then step through donning phases: inflate until the tip senses a resistance drop, hold pressure, and retract the spool for doffing. This suggests a path toward greater autonomy without changing the underlying unfurling mechanism.

The control strategy is therefore notable less for algorithmic complexity than for its minimalism. The device is designed so that morphology and mechanics perform much of the task structure, while pressure acts as the dominant control variable.

5. Experimental characterization and dressing performance

Unfurling-force experiments were conducted with the outer sheath fixed while the subvine everted and pushed against a $F_u$9 load cell. The variables were the number of subvines, $a$0, sheath diameters $a$1, and pressure swept over $a$2–$a$3. The reported results were that $a$4 gives nearly twice the slope of $a$5, validating the global force model; $a$6 suffers slight jamming in a small sheath, which is mitigated by increasing diameter to $a$7; and empirical values of $a$8 from linear fits were $a$9 for $b$0, $b$1 for $b$2, and $b$3 for $b$4 (Kim et al., 9 Jul 2025).

Curvature experiments used a two-rod joint with a bearing simulating an elbow, with one rod free and the other fixed, while SWAG everted through angles from $b$5 to $b$6. Peak subvine pressure and restoring torque were measured. Pressure and torque generally rose with bend angle, but torque plateaued at approximately $b$7, equivalent to lifting $b$8 at $b$9. No buckling occurred up to $f$0, and pressure peaks remained below $f$1.

Dressing demonstrations were reported for three garment configurations:

Configuration	Result
Sleeve	Subject’s arm fully covered in $f$2, posture unconstrained
Jacket	Front-opening jacket closed as channel paths unfurled; donning in $f$3
Pants	Channel routing lifted and sealed an open-front pant garment in $f$4

Across all tests, with $f$5 per configuration, the success rate was $f$6 and the standard deviation in time was $f$7. The summary in the detailed description characterizes the overall performance as fast, approximately $f$8, across garment types, joint angles, and users.

These evaluations support three distinct claims: the pressure–force model captures first-order actuation behavior; the two-subvine architecture can negotiate substantial curvature without buckling; and full dressing demonstrations can be completed without requiring a fixed user pose.

6. Safety, adaptability, and design constraints

A central safety feature is low skin–garment friction. Tip-unfurling ensures that the only sliding interface is between the subvine and its channel, not between the sheath and the skin. The sheath’s TPU side faces outward, maximizing grip on clothing while avoiding slip against itself (Kim et al., 9 Jul 2025).

The sheath is never inflated, so it imparts almost zero normal force on the body except at the tip. Under worst-case bending, the maximum torque measured was approximately $f$9, reported as well below soft-tissue damage thresholds of approximately $F_v = P\,A\,,$0–$F_v = P\,A\,,$1. Operating pressure is capped below the fabric rupture regime, with a pressure limit expressed through

$F_v = P\,A\,,$2

where $F_v = P\,A\,,$3 is set below fabric rupture, approximately $F_v = P\,A\,,$4.

Adaptability arises from garment patterning rather than online perception. Channel geometry can be pre-patterned on any garment, including sleeve, jacket, and pant configurations, and the same subvine mechanism automatically follows body contours, implicitly guided by the user’s limb. No requirement is imposed for the user to hold a fixed pose or to wear specialized sensors.

The bending-compliance constraint is stated as

$F_v = P\,A\,,$5

where $F_v = P\,A\,,$6 ensures that the sheath can traverse a joint of angle $F_v = P\,A\,,$7 with wrinkles that do not stall eversion. This links safety and functionality: insufficient stiffness would compromise force generation, whereas excessive stiffness would hinder joint traversal and adaptability.

Taken together, these features define SWAG as a soft robotic alternative to conventional robotic dressing assistance. Its distinguishing characteristics are tip-first unfurling, pressure-driven growth through subvines, low normal loading, implicit contour following, and a garment-integrated architecture that decouples dressing from rigid manipulation (Kim et al., 9 Jul 2025).