Origami Catapult Mechanisms: Design & Optimization

• Origami catapult mechanisms are innovative deployable structures that use origami folding principles to achieve high energy density and rapid, programmable actuation.
• They leverage specialized crease patterns, including TMP bellows and modular spring joints, to optimize nonlinear force-displacement responses and tunable stiffness.
• Advanced simulation methods, experimental validations, and crease engineering techniques underpin their applications in robotics, deployable microdevices, and dynamic actuators.

Origami catapult mechanisms are functional deployable structures that utilize principles of origami (the art and science of folding flat sheets into three-dimensional configurations) to store and release mechanical energy for throwing, launching, or jumping tasks. These mechanisms leverage well-controlled geometric patterns and nonlinear elasticity to achieve high energy densities, programmable motion amplification, and compact integration within robotic and engineered systems. The rigorous study and experimental characterization of such mechanisms have established their advantages over conventional springs and linkages in applications requiring rapid actuation and tunable compliance.

1. Fundamental Origami Catapult Architectures

Origami catapult mechanisms derive their functionality from the specialized kinematics and energy storage of folded crease patterns. Principal architectures include:

• Strain-Softening Bellows Catapults: Based on the Tachi–Miura Polyhedron (TMP) bellow, these catapults utilize a one-dimensional stack of Miura-ori–derived unit cells with key geometric parameters such as number of cells $N$, panel heights $d$, and sub-fold angle $\alpha$. The TMP bellow exhibits a nonlinear (strain-softening) force-displacement curve, enabling energy storage that exceeds a linear spring with the same peak force and displacement; this behavior is tailorable via crease geometry (Sadeghi et al., 2020).
• Reverse-Fold and Spring-Joint Launchers: Conventional reverse folds offer only a single programmatic parameter (starting angle), whereas “spring joints”—stacked or modular chains of reverse folds and $\pi$-folds—provide nonlinear, highly amplified output motion for a given trigger fold, optimizing angular velocity transfer in throwing arms (Smith, 2024).
• Tunable Bellows Modules: Radially expandable bellows made from stacked polygonal frusta, with stiffness controlled by a single geometric cone angle parameter, and further tunable by nesting multiple bellows in parallel for additive stiffness. These provide high energy density springs that can be stably and repetitively compressed/released in catapult architectures (Chen et al., 2019).

2. Governing Mechanics and Energy Transfer

The mechanics of origami catapult mechanisms are governed by the elastic deformation of crease lines, panel bending, and spatial kinematic amplification.

• TMP Nonlinear Force Law: For a TMP bellow, the axial force is derived (under rigid-panel, torsional-crease assumptions) as:

$$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$

Kinematic relations link fold angles to compression $\Delta y$; the coupled system enables highly programmable nonlinear (softening or stiffening) force-displacement characteristics (Sadeghi et al., 2020).

• Spring Joint Kinematics: The angular amplification of a spring joint composed of $n+1$ reverse folds and $n$ $\pi$-folds is:

$$\phi_{\text{out}}(\xi) = -\pi(n-1) + \sum_{k=0}^n 2\arctan[\cos(\xi/2)\tan(\phi^k_0/2)]$$

where $d$0 are programmable starting angles (Smith, 2024). This enables instantaneous angular velocity gains exceeding 10$d$1 for small trigger motions, limited by self-interference.

• Tunable Stiffness and Force Scaling: For bellows-based catapults,

$d$2

where $d$3 is the cone angle (Chen et al., 2019). Multiple bellows in parallel sum stiffnesses linearly: $d$4 for $d$5 identical modules.

Energy input $d$6 translates to projectile launch height $d$7 via $d$8 (neglecting losses), with tuning possible through crease patterning, angle selection, and modular stacking.

3. Advanced Crease Engineering and Hysteresis Minimization

Mechanical performance is fundamentally affected by the fold-crease implementation:

• PALEO Crease Technology: Plastically annealed lamina emergent origami (PALEO) creates sharp, low-hysteresis hinges using laser-cut perforations and thermal annealing above $d$9. This increases the plasticity threshold, reduces energy loss, and improves folding range. Empirically, efficiency ratios of $\alpha$0 can be obtained in nonlinearly tuned TMP bellows, compared to $\alpha$1 for scored PET creases (Sadeghi et al., 2020).
• Independent Crease Stiffness Control: Distinct tuning of main-fold $\alpha$2 and sub-fold $\alpha$3 stiffness through perforation geometry enables optimization of softening/nonlinearity without compromising global strength. Optimal strain softening occurs when $\alpha$4.
• Layer Management in Spring Joints: For high-amplification mechanisms, modular spring joints circumvent excessive paper layer buildup, critical for practical realization in thick or rigid materials (Smith, 2024).

4. Simulation Methodologies and Computational Design

Numerical modeling and optimization underlie the contemporary design of origami catapult mechanisms:

• Deformable-Body Simulation (MuJoCo): Origami sheets are modeled as graphs $\alpha$5 with panels meshed via constrained triangulation. Panels behave as hyperelastic St Venant–Kirchhoff elements, and creases are implemented as lines of reduced axial stiffness. Dynamic equations incorporate mass, damping, external actuation, and contact constraints to predict physical behavior through time-stepping. The overall elastic potential $\alpha$6 aggregates energy from edges and faces (Han et al., 13 Nov 2025).
• Parameter Optimization (CMA-ES): Simulation-driven design solves for optimal geometric parameters—such as fold angle $\alpha$7 and arm length $\alpha$8—via Covariance Matrix Adaptation Evolution Strategy, maximizing objective functions such as projectile throw distance. Grid searches and evolutionary runs reveal high-performance parameter regions (e.g., $\alpha$9, $\pi$0 mm yielding throws $\pi$1 m), with close correspondence to experimental measurements on 3D-printed prototypes (Han et al., 13 Nov 2025).

5. Experimental Performance and Design Guidelines

Extensive experimental validation and empirical modeling support the practical implementation of origami-based catapult systems:

• Performance Metrics: In TMP-bellow catapults, nonlinear stiffness yields a 9% increase in airtime and 13% higher jump height relative to linear-stick implementations when tested with matched compression and mass ($\pi$2 mm, $\pi$3 g), with statistical significance $\pi$4 and experimental-theoretical disparities within 10–15% (Sadeghi et al., 2020). For simulation-optimized Y-hinge catapults, experimental throw distances of $\pi$5 m for optimized designs closely track simulated predictions (Han et al., 13 Nov 2025).
• Guidelines for Origami Catapult Synthesis:

| Design Dimension | Recommendation | Source | |--------------------------|-----------------------------------------------------------------------|----------------| | Crease Engineering | Use PALEO + high $\pi$6; aim for $\pi$7 | (Sadeghi et al., 2020) | | Geometry (TMP) | $\pi$8–$\pi$9, $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$0 to stay clear of facet collision ($$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$1) | (Sadeghi et al., 2020) | | Spring Joint Amplifier | $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$2–$$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$3, $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$4, modular for thick panels | (Smith, 2024) | | Bellows Stiffness | Tune $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$5–$$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$6 for $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$7–$$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$8 N/m | (Chen et al., 2019) | | Parallel Stacking | Use $$F(\theta_M, \theta_S) = -\frac{32}{d\cos\theta_M} \left[ k_M \frac{N-1}{N}(\theta_M-\theta_{M0}) + k_S \frac{ \cos^3(\theta_G/2)\sin\theta_M }{\cos\alpha\sin\theta_S}(\theta_S-\theta_{S0}) \right].$$9 layers for $\Delta y$0 | (Chen et al., 2019) | | Launch Energy Sizing | $\Delta y$1 | (Chen et al., 2019) | | Latch/Release | $\Delta y$21 ms trigger to minimize energy loss | (Chen et al., 2019) |

• Fatigue and Power Delivery: REBO-style bellows sustain $\Delta y$3 cycles with $\Delta y$4 length drift; energy dissipation is $\Delta y$5–$\Delta y$6 per cycle post-break-in, and single-launch peak power can reach $\Delta y$7 W for a $\Delta y$8 kg load (Chen et al., 2019).

6. Applications and Outlook

The convergence of geometric tunability, nonlinear mechanics, and computational optimization enables origami catapult mechanisms to outperform conventional designs in energetic efficiency, compactness, and adaptability. Their validated applications include robotic jumpers, deployable microdevices, and dynamically dexterous systems. The foundational advances in crease engineering (PALEO), spring-joint design, and rapid design-simulation-optimization pipelines are fostering new categories of programmable, energy-efficient actuators within both research and emerging commercial domains (Sadeghi et al., 2020, Smith, 2024, Han et al., 13 Nov 2025, Chen et al., 2019). Further research directions include high-cycle life testing, material innovation for greater scaling, integration with soft robotic and hybrid systems, and refinement of simulation models for more accurate real-world performance prediction.