Tip-Growing Eversion Robots

• Tip-growing eversion robots are soft continuum systems that extend by everting a tubular membrane, enabling frictionless and precise movement in confined spaces.
• They integrate soft actuators like series pouch motors, cPAMs, and tip mounts to achieve controlled bending, steering, and manipulation in challenging environments.
• Recent innovations incorporate distributed variable stiffness, modular tip payload mounts, and multi-segment designs to enhance maneuverability, retraction, and overall operational reliability.

Tip-growing eversion robots, also termed “vine robots,” are a class of soft continuum robots that achieve mobility through the continuous eversion of an inextensible tubular membrane at their tip. This mechanism, which produces net forward growth without sliding friction along the body or reliance on wall reactions, enables access to remote, cluttered, or tortuous environments with minimal disturbance. Steering, shape control, and tip sensor-payload transport are achieved by integrated soft actuators, selective stiffness modulation, or external tip-mount mechanisms. Recent advances, including underwater hydraulic eversion, multi-segment steering, and distributed variable stiffness, have expanded the operational and application domains of these robots significantly.

1. Tip-Growing Eversion Principle and Body Architecture

Tip-growing eversion robots consist of a thin, soft tube (typically polyethylene, TPU-nylon, or ripstop nylon) stored inverted inside itself at a base. Actuation via internal fluid (air or water) pressure in the main chamber causes the tip to turn inside out, “everting” new material forward as the robot extends. This mechanism ensures that the body does not slide against the environment, minimizing friction and enabling passage through constricted pathways (Coad et al., 2019, Kaleel et al., 2024). Eversion rate is determined by the volumetric inflow $Q$ and cross-sectional area $A$, $v = Q / A$, with practical extension speeds reported from 5–20 mm/s at pressures of 1–60 kPa depending on tube geometry (Al-Dubooni et al., 2024, Agharese et al., 2023).

The everting body can accommodate sensors, tendons, pneumatic actuators, or hydraulic pouches for steering and manipulation. For underwater use, eversion with ambient water yields exact cancellation of buoyant force between internal and displaced fluid, facilitating precise neutral buoyancy (Kaleel et al., 2024).

2. Steering and Maneuverability Mechanisms

Steerability of tip-growing robots is typically achieved by local shortening on one side of the tube, implemented via soft pneumatic (or hydraulic) actuators running longitudinally along the robot:

• Series pouch motors: Welded rectangular pockets that contract under fluid inflation, producing curvature proportional to the difference in arc length between actuated and passive sides (McFarland et al., 26 Oct 2025, Kübler et al., 2023).
• Cylindrical pneumatic artificial muscles (cPAMs): Integrated folded actuators bulging into cylindrical form under pressure, yielding high curvature and low hysteresis (Kübler et al., 2023, Kübler et al., 2022).
• External tip steering mount: Rigid tip modules with tendon-driven 3-DOF joints (e.g. silicone spherical articulators); allow growth-direction changes and active bracing in narrow pipes while decoupling steering from length (Qin et al., 9 Jul 2025).
• Hydraulic bending-pouch steering: Underwater robots use fluid-filled side pouches to produce quasi-static bends; maximum observed bending angles of 68° at 2 L volume input, equivalent to pneumatic land-based systems (Kaleel et al., 2024).

Characteristic curvature metrics ($\kappa$) and steerability thresholds have been systematically quantified. For 3D self-supporting robots, steerability decreases with increasing tip load and is optimized at moderate chamber pressure; integrated actuators require higher pressure ratios but yield greater curvature, while externally attached actuators bend earlier but saturate more quickly (McFarland et al., 26 Oct 2025, Kübler et al., 2023).

3. Distributed Variable Stiffness and Passive Navigation

To enable complex shape control and buckling prevention, tip-growing robots may incorporate local or distributed variable-stiffness modules. Two principal approaches have emerged:

• Positive pressure layer jamming: Circumferential pouches with layered frictional sheets (e.g., paper) are compressed via regulated pouch pressure; jammed segments can increase bending stiffness by up to an order of magnitude, allowing virtual joint creation and multi-segment shape locking with minimal DOAs (Do et al., 2023, Do et al., 2020).
• Passive constrictive bands: Inextensible circumferential TPU bands welded along the body at regular intervals induce local wrinkling and dramatic drops in flexural rigidity (up to 91%), enabling reliable traversal of high-curvature (25 mm bend radius) features with minimal complexity (Suulker et al., 18 Jan 2026).

The underlying mechanics are accurately captured by Cosserat rod models parameterized by local diameter and effective modulus, with validated predictive value in both rigid and compliant phantoms (Suulker et al., 18 Jan 2026, Do et al., 2023).

4. Tip Payload Mounting and Sensor Integration

Maintaining a sensor or tool at the tip during growth and retraction, while preserving eversion ability, is a nontrivial engineering challenge:

• Mechanically interlocked tip mounts: Modular tip devices (PLA cap, roller retraction motor, magnetic rolling interlock) follow the everting wall automatically during both growth and retraction without slip or unwanted detachment, supporting multi-kg payloads and controlled sample retrieval (Jeong et al., 2019, Coad et al., 2019).
• Deformable textile caps: Fabric-based slip-on caps (elastic knit or banded skirt) achieve robust frictional holding while accommodating radial compression for tight passage or protrusion clearance, supporting camera transport and squeeze-through operations (Suulker et al., 2024, Suulker et al., 2023).
• Selective multi-layer soft mounting: Multi-material layered caps with distributed surface friction maintain payload orientation and position across multi-layer and feature-rich environments (Suulker et al., 2023).

Sensor systems typically include embedded pressure sensors for air-pocket contact detection (Mitchell et al., 2023), tip-mounted IMUs for 3D odometry (Qin et al., 9 Jul 2025), and passive or externally driven force-sensing via compliant skin.

5. Retraction, Shape Locking, and Workspace Extension

Retraction without buckling is enabled by tip-mounted roller devices that capture and invert the wall at the tip rather than the base, nullifying Euler-type buckling failures for arbitrary lengths. Analytical models relate tip tension and operating pressure to buckling thresholds, governing safe retraction profiles (Coad et al., 2019). This mechanism enables true reversible exploration and manipulation in tortuous or branched conduits.

Shape-locking is achieved via tip-extending, pressurized side chambers constrained by rigid guides, allowing selective freezing of curvature for workspace expansion and dexterity improvement by concatenated constant curvature segments; simulations and experiments report reachable area increases by up to 300% over non-locking robots (Wang et al., 2020).

6. Advanced Architectures: Multi-Segment Robots and Preformed Pathways

Multi-segment tip-growing robots implement selective, segment-wise actuation via tip-mounted magnetic valve actuators, enabling piecewise constant curvature paths and controlled, repeatable 3D navigation without environmental contact (Kübler et al., 2022). Preformed robots fabricated with discrete folds, welds, or loop-and-screw connectors achieve high-fidelity shape programming along designated paths; measured joint-angle errors as low as 0.12° and growth speeds from 3.8–10 cm/s, with fabrication choice monotonically affecting mechanical performance (Agharese et al., 2023).

Hybrid vine-continuum robots integrate tendon-driven tip caps with soft growing bodies for combined long-reach access and precise end effector steering, supporting high-fidelity manipulation and targeted payload delivery (e.g., decontamination in nuclear vessels) (Al-Dubooni et al., 2024).

7. Operational Limits, Design Trade-Offs, and Future Directions

Material selection, actuator geometry, payload mass, chamber pressure, and actuator integration method all critically affect steerability, shape retention, and growth rate. Optimized designs require balancing tip load for curvature, choosing diameter to stay above collapse threshold, and coordinating actuator pressures for rapid growth and sharp turning (McFarland et al., 26 Oct 2025, Kübler et al., 2023). Underwater operation introduces additional constraints on sealing, leak management, and dynamic buoyancy control (Kaleel et al., 2024).

Limitations include prototype variability due to manual fabrication, open-loop volume control leading to repeatability issues, and challenges in wiring payloads through soft tips. Emerging directions include automated heat-sealing and multi-material films for improved durability, closed-loop vision/IMU feedback for high-precision steering, miniaturized external tip mounts, and generalization to medical navigation, search-and-rescue, and confined-environment inspection (Suulker et al., 2024, Suulker et al., 18 Jan 2026, Al-Dubooni et al., 2024).

---

In summary, tip-growing eversion robots combine scalable, frictionless locomotion with a growing array of soft actuation, steering, variable-stiffness, and tip-payload technologies. These innovations are underpinned by robust mathematical models, extensive benchmarking, and validated deployment in complex field and laboratory environments (Kaleel et al., 2024, Stroppa et al., 2019, Qin et al., 9 Jul 2025, Do et al., 2020, Kübler et al., 2023, Agharese et al., 2023, Suulker et al., 18 Jan 2026, Kübler et al., 2022, Jeong et al., 2019).