Enabling High-Curvature Navigation in Eversion Robots through Buckle-Inducing Constrictive Bands

TL;DR

This work tackles the challenge of navigating high-curvature environments with tip-growing eversion robots by introducing passive constrictive bands that locally reduce bending stiffness. Modeling the robot as a Cosserat rod with diameter-reducing bands, the authors show that stiffness can be significantly lowered without adding active actuators, enabling tighter bends and smoother navigation. Experimental validation across multiple configurations demonstrates substantial stiffness reductions (up to ~91% at the tip) and improved navigation in rigid and colon-like environments, albeit with a trade-off in higher eversion pressures for larger diameter reductions. The approach preserves softness and low mechanical complexity, expanding the practical reach of eversion robots in medical and inspection contexts.

Abstract

Tip-growing eversion robots are renowned for their ability to access remote spaces through narrow passages. However, achieving reliable navigation remains a significant challenge. Existing solutions often rely on artificial muscles integrated into the robot body or active tip-steering mechanisms. While effective, these additions introduce structural complexity and compromise the defining advantages of eversion robots: their inherent softness and compliance. In this paper, we propose a passive approach to reduce bending stiffness by purposefully introducing buckling points along the robot's outer wall. We achieve this by integrating inextensible diameter-reducing circumferential bands at regular intervals along the robot body facilitating forward motion through tortuous, obstacle cluttered paths. Rather than relying on active steering, our approach leverages the robot's natural interaction with the environment, allowing for smooth, compliant navigation. We present a Cosserat rod-based mathematical model to quantify this behavior, capturing the local stiffness reductions caused by the constricting bands and their impact on global bending mechanics. Experimental results demonstrate that these bands reduce the robot's stiffness when bent at the tip by up to 91 percent, enabling consistent traversal of 180 degree bends with a bending radius of as low as 25 mm-notably lower than the 35 mm achievable by standard eversion robots under identical conditions. The feasibility of the proposed method is further demonstrated through a case study in a colon phantom. By significantly improving maneuverability without sacrificing softness or increasing mechanical complexity, this approach expands the applicability of eversion robots in highly curved pathways, whether in relation to pipe inspection or medical procedures such as colonoscopy.

Figures (5)

• Figure 1: a) Concept drawing of an eversion robot equipped with constrictive bands navigating through a colon-like environment. b) Illustration of a standard eversion robot getting stuck at a sharp bend. c) Illustration of the proposed eversion robot with integrated constrictive bands successfully navigating the same bend. d) Detailed view of an eversion robot with the constrictive bands incorporated.
• Figure 2: Outline of the fabrication process with constrictive band implementation. a) Overview of the eversion robot with constrictive bands. b) The eversion robot is manufactured using a substrate sheet while each constraining band is made of a further piece of the same TPU material. c) The ends of each constraining piece are sealed to the left and right edges of the substrate sheet, while three intermediate points are also affixed to ensure uniform contraction and circular symmetry during eversion. Sealing is achieved using an ultrasonic welding machine. d) Once the band strips are fixed into place, the sheet is carefully folded lengthwise to align its edges, creating a tubular form. e) The edges are then sealed along the seam to create an airtight cylindrical structure, later to be inverted to create an eversion robot.
• Figure 3: Experimental configurations and results for characterization experiments 1–3. In all the experiments, the eversion robot was displaced laterally from the tip by 20 mm. The mathematical simulation results are scattered on plots. a) In Experiment 1, the number of constrictive bands was varied, and the corresponding results are shown in b). c) In Experiment 2, the band position was altered, and these results are shown in d). e) In Experiment 3, the diameter reduction rate was varied, and these results are presented in f).
• Figure 4: Top - Experimental setup and results for Characterization Experiment 4. a) The eversion robot was inflated in a controlled environment, and the minimum pressure required for full eversion was recorded. b) Results indicate a super-linear increase in the required eversion pressure as the robot diameter is increasingly restricted by the bands. Bottom - Experimental setup and results for the Validation Experiment in a confined environment. c) The robot is fixed inside a tube and inflated at 3 kPa. The distal end of the tube has 180$\degree$ turn casing attachments of progressively decreasing bending radii. The trial is deemed a success if the robot everts through the bend, and a failure if it gets stuck within it. d) The results demonstrate that the robot with 10% constrictive bands achieves the highest overall performance in this validation experiment.
• Figure 5: Case study using a colon phantom. a) A silicone phantom fixed to a whiteboard with elastic bands and magnets; the eversion robot is inserted using a funnel. b) The standard eversion robot stalls at the 90° bend. c) The band-equipped robot successfully navigates the same bend and fully everts. d) Pressure–position plot showing that the standard robot gets stuck before the 400 mm mark, even at the maximum actuation pressure of 8 kPa, while the band-equipped robot completes the bend at a maximum actuation pressure of just over 6 kPa.