Modeling Collapse of Steered Vine Robots Under Their Own Weight

TL;DR

The paper advances vine-robot stability analysis by introducing a tail-tension-aware, true-shape collapse model that unifies predictions for both steered and unsteered configurations. It introduces a discretized, segment-based approach to incorporate real robot shapes and quantifies the evert force F_e, enabling accurate collapse predictions across diameters, pressures, and steering methods, including inflated supports. Extensive experimental validation covers unsteered straight pipes, inflated supports, and steered configurations, with demonstrations showing field-relevant task performance such as gap crossing. The framework supports real-time shape sensing and could underpin design and planning tools for 3D navigation tasks, while highlighting future work on dynamics, material choices, and tail-movement effects.

Abstract

Soft, vine-inspired growing robots that move by eversion are highly mobile in confined environments, but, when faced with gaps in the environment, they may collapse under their own weight while navigating a desired path. In this work, we present a comprehensive collapse model that can predict the collapse length of steered robots in any shape using true shape information and tail tension. We validate this model by collapsing several unsteered robots without true shape information. The model accurately predicts the trends of those experiments. We then attempt to collapse a robot steered with a single actuator at different orientations. Our models accurately predict collapse when it occurs. Finally, we demonstrate how this could be used in the field by having a robot attempt a gap-crossing task with and without inflating its actuators. The robot needs its actuators inflated to cross the gap without collapsing, which our model supports. Our model has been specifically tested on straight and series pouch motor-actuated robots made of non-stretchable material, but it could be applied to other robot variations. This work enables us to model the robot's collapse behavior in any open environment and understand the parameters it needs to succeed in 3D navigation tasks.

Figures (13)

• Figure 1: Demonstration where our steered vine robot collapse model successfully predicts a lack of collapse during a gap-crossing task. (a) The robot curves to avoid deflecting off course. (b) The curve enables the robot to grow high enough such that when it deflects, it lands on the other side of the gap instead of deflecting underneath it or collapsing.
• Figure 2: Diagram of vine robot parameters used in our collapse model (modified from mcfarland2023collapse). (a) The robot consists of two layers of material with length $L$, diameter $D$, and thickness $t$. (b) The robot, growing at an angle $\gamma$ above the horizontal, has three forces acting on it in the directions shown: gravity pointing down ($mg$), tail tension pointing back toward the base ($T_t$), and air pressure ($P$) pointing in the direction of growth. Air pressure acts on all walls of the robot, but the net force points toward the tip. The center of mass is a distance $r$ from the base of the robot.
• Figure 3: Plot of expected tail tension values varying with pressure for a quasistatic vine robot as used in our experiments. For a quasistatic robot, tail tension is approximately half of the robot's internal body pressure $P$, multiplied by its area $A$. However, uncertainty about how the pressure force is shared between the robot's tail and wall can cause the tail tension to be higher or lower than that average value. The maximum amount by which a quasistatic robot will deviate from that value is$\frac{F_e}{2}$, that is, the force needed to evert the robot. The purple curve shown represents the point at which the robot begins inverting. The yellow curve represents the point at which the robot begins everting or growing, and we expect the tail tension value for our robot to be approximately along this line. In between the two boundary curves, the robot is stopped. The red curve represents the average value of tail tension, which is between the two extremes. It is possible to achieve values outside the boundary curves by inverting or everting quickly.
• Figure 4: Experimental setup for positioning vine robots at various angles, used to obtain the results shown in Figure \ref{['giant summary plot']} (reprinted from mcfarland2023collapse). (a) The robot is pushed out of the pipe, which is positioned at some adjustable growth angle $\gamma$ above horizontal. (b) If the robot collapses due to transverse loading, it will do so at the edge of the pipe, which is the last point of support.
• Figure 5: Results of how collapse length varies with growth angle, pressure, and diameter for unsteered fabric robots. (a) Collapse length vs. growth angle plot for a 2.43 cm diameter robot at 3.45 kPa (modified from mcfarland2023collapse). Collapse length vs. pressure plots for a 2.43 cm diameter robot at growth angles of (b) 0$^{\circ}$ and (c) 45$^{\circ}$. Collapse length vs. diameter plots for (d) a growth angle of 0$^{\circ}$ and a pressure of 2.07 kPa, (e) a growth angle of 0$^{\circ}$ and a pressure of 6.89 kPa, and (f) a growth angle of 45$^{\circ}$ and a pressure of 2.07 kPa. The predicted collapse lengths for our model incorporating no tail tension (blue solid line), average tail tension (red dashed line), eversion tail tension (yellow dotted line), and inversion tail tension (purple dash dot line) are shown along with the data (black circles). The eversion tail tension model has the best match with all six configurations. The latter three models mentioned above are also shown in Figure \ref{['supports plot']}.