💻 computer science

Design and Characterization of a Dual-DOF Soft Shoulder Exosuit with Volume-Optimized Pneumatic Actuator

This paper presents a lightweight, dual-degree-of-freedom soft shoulder exosuit featuring a volume-optimized spindle-shaped angled actuator that significantly improves dynamic response while maintaining torque output, ultimately demonstrating substantial reductions in user muscle activity during shoulder abduction and flexion tasks.

Original authors: Rui Chen, Domenico Chiaradia, Daniele Leonardis, Antonio Frisoli

Published 2026-02-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Rui Chen, Domenico Chiaradia, Daniele Leonardis, Antonio Frisoli

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine your shoulder is like a complex, high-performance hinge that lets you reach for the sky, hug a friend, or grab a coffee cup. Now, imagine that hinge gets stiff or weak due to injury or aging. You need a "muscle helper" to take some of the load, but you don't want a clunky, heavy metal robot arm strapped to you. You want something light, flexible, and invisible until you need it.

This paper is about building exactly that: a soft, wearable "shirt" that acts like a super-powered, invisible muscle to help your shoulder move.

Here is the breakdown of their invention, explained with some everyday analogies:

1. The Problem: The "Heavy Backpack" Dilemma

Most robotic helpers for shoulders are either:

Too heavy: Like wearing a backpack full of bricks (rigid metal robots). They are strong but uncomfortable and restrict your natural movement.
Too weak or slow: Soft fabric robots are comfortable, but to make them strong enough to lift your arm, you usually need a big air pump and a large tank of air. This makes the system bulky, like dragging a scuba tank around.

The researchers faced a tricky trade-off: If you make the air muscle bigger to get more strength, it gets slower and needs more air. If you make it smaller to save air, it loses strength.

2. The Breakthrough: The "Spindle-Shaped" Air Muscle

The team invented a new shape for the air muscle (actuator).

The Old Way: Imagine a cylindrical balloon (like a soda can). It has the same width from top to bottom. To fill it, you need a lot of air, and it takes time to inflate.
The New Way (SSAA): Imagine a spindle or a football (or even a hot dog bun that is pinched at the ends). It is wide in the middle where it needs to push hard, but it tapers off at the ends where it doesn't need to do much work.

The Magic Result: By pinching the ends, they reduced the amount of air needed by 35%.

Analogy: It's like switching from a wide, flat tire to a skinny racing tire. You use less rubber (air), but you can still drive just as fast.
Performance: This new shape inflated 35% faster and used less air, while still keeping 94% of the lifting power. It's the "Goldilocks" zone: just the right size for a portable system.

3. The Suit: A "Smart T-Shirt" with Two Superpowers

They built a soft exosuit (a wearable shirt) that weighs only 390 grams (about the weight of a large apple). It has two main "muscles" sewn into it:

The "Y-Shaped" Lifter (Abduction): This helps you lift your arm straight out to the side (like a bird flapping wings). It's shaped like a "Y" to hug your armpit and arm perfectly, preventing it from sliding around.
The "Pouch" Pusher (Adduction): This helps you push your arm forward and across your chest (like hugging yourself or reaching for a seatbelt). It uses stacked fabric pockets that expand like a bellows.

The Cool Feature: The suit is "modular." If you just need to lift your arm, only one muscle turns on. If you need to reach forward, both turn on. It's like having a car with two different gears for different terrains.

4. The Test: Does it Actually Help?

They tested this on 10 healthy people (not patients yet) to see how well it worked. They measured muscle activity using sensors (like listening to the electrical "hum" of your muscles).

Lifting to the side (Abduction): The suit was a huge success. It reduced muscle effort by up to 59%. It was like the suit was doing half the work for you.
Reaching forward (Flexion): The suit also helped a lot, reducing effort by up to 63%.
The Twist: When they tested both muscles working together for reaching forward, the extra help wasn't huge for healthy people.
- Why? Healthy brains are smart. When the suit pushed too hard, the healthy people's brains said, "Whoa, that's weird," and they tightened other muscles to stabilize themselves. It's like when you try to walk on a moving walkway; you might tense up because you aren't used to it.
- The Hope: The researchers believe that for people with injuries (like stroke survivors) whose brains can't compensate as well, this dual-motor system will be a game-changer because they won't fight the extra help.

5. The Bottom Line

This paper isn't just about a cool shirt; it's about efficiency.

Before: To get a strong, fast robotic shoulder, you needed a heavy, bulky air tank.
Now: By reshaping the air muscles (the "spindle" idea), they made a system that is light, fast, and strong enough to be carried around in a small backpack.

In simple terms: They figured out how to make a "super muscle" out of fabric that is so efficient it doesn't need a giant air tank to work. This brings us one big step closer to robots that can help people with shoulder injuries move freely again, without looking like they are wearing a sci-fi costume.

1. Problem Statement

Portable pneumatic soft exosuits for the shoulder face a critical trade-off between torque output and dynamic response.

System Constraints: Portable pneumatic systems rely on small compressors and reservoirs. Adding multiple actuators (required for complex 2-DOF shoulder movement like abduction and flexion) increases air consumption, reducing the number of achievable repetitions.
Design Limitations: Existing solutions often use uniform cylindrical actuators. To increase torque, actuators must be larger (increasing volume and air demand), which slows down the dynamic response (rise/fall time). Conversely, smaller actuators respond faster but provide insufficient torque.
Gap: There is a lack of dual-DOF soft shoulder exosuits that effectively balance these constraints while providing assistance for both shoulder abduction and flexion/horizontal adduction.

2. Methodology

A. Actuator Design: Spindle-Shaped Angled Actuator (SSAA)

The authors proposed a geometric optimization to resolve the volume-torque-response trade-off.

Concept: Unlike traditional Uniform Cylindrical Angled Actuators (UCAA) with constant cross-sections, the SSAA features a non-uniform cross-section that tapers toward both ends (spindle shape).
Rationale: In bending actuators, the force required at the distal ends (farther from the pivot) is lower than at the center. By reducing the cross-sectional area at the ends, the total air volume is reduced without significantly compromising the moment arm or peak torque generation in the operational range.
Geometry: The SSAA maintains the same middle width ( $w_2$ ) as the UCAA but reduces the upper ( $w_1$ ) and bottom ( $w_3$ ) widths.

B. System Integration

The SSAA geometry was integrated into a dual-DOF textile-based exosuit (total weight: 390 g):

Curved Abduction Actuator (CAA): A Y-shaped actuator constructed from two SSAA units and constraint layers. It assists shoulder abduction.
Horizontal Adduction Actuator (HAA): Based on stacked fabric pouch motors, assisting in horizontal adduction and flexion.

Operation Modes:
- Abduction: CAA only.
- Flexion: CAA + HAA (simultaneous actuation).
- Horizontal Adduction: HAA actuated after CAA.

C. Characterization & Modeling

Mechanical Testing: The SSAA and CAA were tested for moment-angle characteristics, frequency response (Bode plots), and durability (2000+ cycles).
Analytical Modeling: A simplified analytical model was developed for the HAA (pouch motor) to predict force-height and torque-angle relationships, validated against experimental data.

D. User Study

Participants: 10 healthy subjects (5 male, 5 female).
Protocol: Dynamic tasks involving shoulder flexion and abduction under three conditions:
1. No exosuit (Baseline).
2. Single-actuator (CAA only for abduction; CAA+HAA for flexion).
3. Dual-actuator (CAA + HAA).
Metrics: Surface Electromyography (sEMG) recorded from 7 muscles (Biceps, Supraspinatus, Trapezius, Pectoralis Major, Anterior/Posterior/Middle Deltoids). Range of Motion (ROM) and subjective fatigue/comfort questionnaires were also collected.

3. Key Contributions

Volume-Optimized Actuator Design (C1):
- Introduced the SSAA, achieving a 35.7% reduction in volume (357 mL vs. 555 mL) compared to uniform cylindrical designs.
- Maintained 94.2% of the output torque in the operational range ( $\theta > 90^\circ$ ).
- Achieved a 35.2% faster dynamic response (reduced rise/fall time) due to lower air volume requirements.
Dual-DOF System Integration (C2):
- Developed and validated a lightweight (390 g) textile exosuit capable of multi-modal assistance (abduction, flexion, horizontal adduction).
- Provided analytical models for the HAA and experimental validation of the CAA's torque-angle and frequency characteristics.
Empirical Characterization (C3):
- Conducted a systematic sEMG evaluation across seven muscles.
- Quantified the incremental benefit of adding a second actuator (HAA) to a single-actuator system in healthy users.

4. Results

Actuator Performance

Torque: The CAA (using SSAA) achieved a peak moment of 9.7 Nm at 90 kPa and $90^\circ$ , comparable to prior state-of-the-art systems but with significantly less volume.
Response: The SSAA showed a 20–30% reduction in rise time and up to 35.2% reduction in fall time compared to the UCAA.
Durability: No degradation in output force was observed after 2000 inflation/deflation cycles.

User Study Findings

Shoulder Abduction:
- The exosuit (CAA only) significantly reduced muscle activity across all 7 muscles.
- Maximum Reduction: Up to 59% reduction in Biceps Brachii activity.
- ROM was minimally affected (1.9% reduction), which was statistically significant but practically negligible.
Shoulder Flexion:
- Both single-actuator (HAA only, though the text implies CAA+HAA for flexion in the abstract, the results section clarifies the comparison between CAA-only vs. CAA+HAA) and dual-actuator configurations reduced EMG activity by up to 63.7%.
- Incremental Benefit: Adding the HAA to the CAA provided limited additional benefit for healthy users during flexion. Statistically significant additional reductions were observed only in the Pectoralis Major. Other muscles (e.g., Supraspinatus, Upper Trapezius) showed increased activity or no significant change, suggesting compensatory co-contraction to stabilize the joint against the new force vector.
Subjective Feedback: Users reported high comfort, ease of donning/doffing, and significant fatigue reduction with assistance.

5. Significance and Implications

Design Guidance: The study proves that non-uniform cross-sections (SSAA) are a viable strategy to optimize portable pneumatic systems, allowing for faster response times without sacrificing torque. This is crucial for battery-powered, portable rehabilitation devices.
Clinical Insight: The finding that dual-DOF assistance offers limited incremental benefit over single-DOF assistance in healthy users suggests that healthy neuromuscular control may trigger compensatory mechanisms (co-contraction) that negate the benefits of extra degrees of freedom.
Target Population: The authors hypothesize that patients with motor impairments (e.g., stroke, SCI) who lack the ability to compensate may benefit significantly more from dual-DOF assistance. This highlights the need for future validation in clinical populations.
Future Control Strategies: The results suggest that task-aware activation (activating the HAA only during specific phases like horizontal adduction) might be more effective than continuous simultaneous actuation to avoid triggering stabilizer co-contraction.

In conclusion, this work provides a robust design methodology for lightweight, responsive soft exosuits and offers critical data on how multi-DOF assistance interacts with human neuromuscular control, guiding the development of next-generation rehabilitation robots.