Antagonistic Fabric Pneumatic Actuators

• AFPAs are fabric-based soft actuators with pneumatic chambers arranged in antagonistic pairs for bidirectional force control.
• They employ thermoplastic-coated and silicone-impregnated textiles for customizable stiffness, high strength-to-weight ratios, and low-profile integration.
• Advanced control strategies, including machine learning and finite element modeling, optimize AFPAs for use in wearable robotics and haptic interfaces.

Antagonistic Fabric-Based Pneumatic Actuators (AFPAs) are a class of soft robotic actuators employing textile architectures and pneumatic principles to enable bidirectional, variable-stiffness, and multi-modal force generation for wearable robotics, haptic interfaces, and soft manipulators. By leveraging inherently compliant, lightweight, and customizable fabric materials—frequently thermoplastic polyurethane (TPU)-coated textiles or silicone-impregnated nylon—AFPAs achieve advantageous strength-to-weight ratios, low-profile geometry, and high mechanical transparency suitable for integration into exosuits and interactive devices.

1. Design Principles and Materials

The foundational structure of AFPAs consists of two—or more—fabric pneumatic chambers positioned in mechanical opposition such that force output from one (“agonist”) is counteracted by the other (“antagonist”) (Prattico et al., 2014, Schaffer et al., 7 Mar 2024, Chen et al., 5 Sep 2025). These textile chambers are typically fabricated through heat-sealing, glue bonding, or thermal pressing of coated fabrics. Critical fabrication parameters include:

• Material selection: Thermoplastic-coated nylon and ripstop nylon fabrics are favored for their high stiffness-to-weight ratio, airtightness, and gluing compatibility (Coram et al., 1 Nov 2024). TPU-coated textiles, bonded at 260°F for 90 seconds and 5 MPa, yield robust, airtight seams with minimal material drift, maintaining actuator strength over extended use.
• Geometry and modularity: AFPAs adopt modular cell architectures—single-cell, multi-cell, or bellow-type—enabling tailored force profiles and displacement characteristics. The cell number, width, and shape directly influence peak force, range of motion (ROM), and motion smoothness (Sahin et al., 2022, Sahin et al., 2023).

Volume-constrained designs such as “Volume Transfer” relocate actuation volume inward, yielding slim exosuit profiles and enlarged distributed stress areas without compromising torque (Liu et al., 11 Jan 2024). This adjustment enhances wearability for integration beneath clothing, reduces local pressure, and preserves actuator performance.

2. Mechanics and Modeling

AFPAs operate via controlled internal pressurization, actuating fabric structures to generate force and torque at target joints or interaction points. The force output is generally modeled by:

$F = p \cdot A$

where $p$ is internal pressure and $A$ is the active area, with geometry and constraints modulating $A$ during actuation (Ayazi et al., 24 Mar 2025, Sahin et al., 2022). For torque about a joint:

$\tau = F \cdot d$

where $d$ is the moment arm (distance from joint center to force application). Optimization of $d$, via judicious actuator anchoring (e.g., at two-thirds limb length for pediatric exosuits), simultaneously enhances ROM and reduces peak contact forces (Ayazi et al., 17 Jul 2024, Ayazi et al., 24 Mar 2025).

Nonlinearities and hysteresis are observed in force–pressure relationships due to geometric transitions (e.g., rapid unfolding of actuators) and are experimentally mapped with embedded load cells and encoders (Ayazi et al., 17 Jul 2024). Piecewise models such as:

$$F(P) = \begin{cases} \alpha P, & P < P_c \ \beta P, & P \geq P_c \end{cases}$$

capture the abrupt shift in contact area at critical pressures $P_c$ (Ayazi et al., 17 Jul 2024).

Finite element (FE) frameworks incorporating anisotropic, hyperelastic material constitutive equations and geometry-based contraction models, as well as dynamic explicit solvers ($M \cdot \ddot{u} + D \cdot \dot{u} + K \cdot u = F$), are used to predict buckling, stress distributions, and optimize mechanical performance for specific textile and anchoring modalities (Pasquier et al., 2023).

3. Control Strategies

Classical AFPAs employ a “couple control” paradigm, where joint torques are computed via Newton–Euler dynamics and mapped, using actuator geometry, to desired pressures via an inverted polynomial fit:

$$f(x,y) = a_1 + a_2 x + a_3 y + a_4 x^2 + a_5 x y + a_6 y^2 + a_7 x^2 y + a_8 x y^2 + a_9 y^3$$

$x$ being supply pressure, $y$ being contraction (Prattico et al., 2014). The inversion yields supply pressures required for the computed forces at known contraction.

Recent advances utilize machine learning approaches—Gaussian processes (GPs) trained on input-output data—to learn the nonlinear mapping from desired joint position and stiffness to actuator pressures, bypassing explicit model dependencies and accommodating material nonlinearities and system hysteresis (Habich et al., 2023). Feedforward GP predictions combined with low-gain PI feedback achieve position errors as low as $0.34^\circ$ and compensate typical pressure errors of 11.5% full range.

Dual-chamber pressure regulation architectures (“HapMorph”) enable independent modulation of multiple mechanical properties (e.g., geometry and stiffness), formalized as:

$$\delta W = P_1 \delta V_1 + P_2 \delta V_2 - F \delta H = 0$$

$$F = P_1 \left(\frac{\partial V_1}{\partial H}\right) + P_2 \left(\frac{\partial V_2}{\partial H}\right)$$

where $P_1, P_2$ are pressures in the modulating and morphing chambers, and $\frac{\partial V_x}{\partial H}$ is virtual work derivative with respect to height $H$. This permits mechanical decoupling of size and stiffness in wearable haptic interfaces (Chen et al., 5 Sep 2025).

4. Performance Characterization

AFPAs deliver rapid dynamic response, high durability, and robust force output in wearable formats:

• Force output and efficiency: Blocked force measurements indicate textile actuators produce up to 36.1 N (95.3% of elastomeric counterparts), with a profile reduced by 96.4% in thickness and 57.2% in mass (Coram et al., 1 Nov 2024).
• Dynamic actuation: Rise times as low as 2.12 s (modular sleeve shoulder actuators) and symmetric inflation–deflation cycles (<2 s for certain designs) support responsive control in real-time exosuits (Natividad et al., 2019).
• ROM and smoothness: Optimal actuator placements (e.g., two-thirds upper arm for shoulder abduction) maximize ROM while minimizing contact forces; multi-cell and larger-width designs provide greater motion while single-cell variants deliver higher peak force but with increased variability (Sahin et al., 2022, Sahin et al., 2023, Ayazi et al., 17 Jul 2024).
• Consistency and resilience: Cyclical testing over 400 inflation cycles reveals textile actuators maintain stable output force profiles, outperforming elastomeric actuators in drift resistance (Coram et al., 1 Nov 2024).

5. Application Domains

AFPAs are implemented across assistive robotics, wearable haptic technology, and soft manipulation:

• Exosuits for rehabilitation: Pediatric upper-extremity exosuits employ bidirectional and multi-cell fabric actuators for shoulder abduction/adduction and elbow extension/flexion, with experimental findings establishing best practices for anchoring points and joint configurations to optimize function and user comfort (Ayazi et al., 17 Jul 2024, Sahin et al., 2023, Ayazi et al., 24 Mar 2025).
• Haptic interfaces: HapMorph demonstrates rendering of nine distinct tactile states (three size categories × three stiffness levels) with 89.4% participant discernment accuracy (Chen et al., 5 Sep 2025). Textile actuators integrated in glove, sleeve, and vest architectures can simulate geometric and mechanical properties, augmenting VR/AR experiences and social communication.
• Soft robotic grippers and manipulators: Star-shaped gripper architectures capture in-plane overcurvature mechanics to deliver high stiffness-to-weight ratios; stacking and modular arrangements further scale force output and object adaptability (Andrade-Silva et al., 2022).
• Exosuit anchoring and comfort augmentation: fPAM sleeves provide adjustable, pressure-regulated anchoring mechanisms with low mounting point displacement, high burst pressure limits, and inherent pre-load capacity when uninflated (Schaffer et al., 7 Mar 2024).

6. Design Tradeoffs and Limitations

Key tradeoffs in AFPA implementation include:

• Force–pressure–profile optimization: Maximizing torque and range of motion while maintaining wearability and low-profile designs may require iterative FE simulation and dynamic testing, especially when applying volume transfer principles (Liu et al., 11 Jan 2024).
• Smoothness vs. range of motion: Increasing actuator cell count or size enhances angular ROM but may reduce straightness (smoothness) and inflate/deflate times—posing design constraints for time-critical movements in pediatric devices (Sahin et al., 2023).
• Material variability and fabrication consistency: Air-tightness and seam strength are highly sensitive to fabrication parameters (temperature, pressure, time); sub-optimal processing leads to increased air leakage and abrupt reduction in force capacity (Coram et al., 1 Nov 2024).

7. Future Directions

Research in AFPAs is progressing toward:

• Online learning and adaptive control: Continually adapting GP models to account for material fatigue and operational lifecycle (Habich et al., 2023).
• Advanced FE-based parametric optimization: Open-source simulation frameworks enable rapid iteration on material, seam, and geometry design to maximize stress management and functional performance (Pasquier et al., 2023).
• Multi-modal haptic rendering: Extension of antagonistic architectures to texture, temperature, and multimodal feedback via integration with additional sensory and actuation systems (Chen et al., 5 Sep 2025).
• Enhanced wearability and modularity: Further refinement of anchoring systems, actuator volume distribution, and personalized torque optimization offers promise for widespread adoption in daily-wear applications (Liu et al., 11 Jan 2024, Schaffer et al., 7 Mar 2024).

AFPAs, by leveraging the interplay of fabric mechanics, pneumatic principles, antagonistic arrangement, and advanced control methodologies, represent a significant advancement in the field of soft wearable robotics and haptic technology, enabling fine-tuned, robust bi-directional actuation and multi-property rendering within practical human-centered constraints.