On Steerability Factors for Growing Vine Robots

Basic Information

• Paper ID: 2510.22504
• Title: On Steerability Factors for Growing Vine Robots
• Authors: Ciera McFarland, Antonio Alvarez, Sarah Taher, Nathaniel Hanson, Margaret McGuinness
• Classification: cs.RO (Robotics)
• Institutions: University of Notre Dame, MIT Lincoln Laboratory
• Publication Date: October 26, 2025 (arXiv submission)
• Paper Link: https://arxiv.org/abs/2510.22504

Abstract

Vine robots extend their tubular bodies by everting material from the tip, enabling navigation through complex environments with minimal soft-body structure. Despite promising applications in urban search and rescue, their performance is constrained by sensor/tool weight and design and control parameters. This study systematically investigates how tip load, pressure, length, diameter, and manufacturing method affect vine robot steerability—the ability to maneuver with controlled curvature. The research comprises two experimental sets: (1) investigating tip load, chamber pressure, length, and diameter under self-supporting conditions against gravity; (2) investigating manufacturing method and actuator-to-chamber pressure ratio under ground-supported conditions. Results demonstrate that steerability decreases with increasing tip load, is optimal at moderate chamber pressure, improves with length, and is essentially unaffected by diameter. External actuator robots begin bending at low pressure ratios but saturate at high ratios; integrated actuator robots require higher pressure ratios to bend but achieve higher overall curvature.

Research Background and Motivation

Core Problem

Vine robots, through their unique tip-eversion growth mechanism, can explore confined spaces and have demonstrated application potential in urban search and rescue training sites, archaeological sites, and salamander cave habitats. However, existing deployments rely primarily on ad hoc design and control parameters, lacking systematic guiding principles.

Problem Significance

1. Practical Application Needs: Robots must carry sensors and tools into hazardous environments, where tip load significantly impacts maneuverability
2. Performance Limitations: Robot mobility is constrained when traversing open spaces or carrying tip loads
3. Design Complexity: Tip mass, structural requirements, and control requirements are mutually coupled, requiring systematic quantification

Limitations of Existing Approaches

• Prior research has characterized unloaded curvature, growth speed, and turning force, but lacks systematic study of 3D steerability under load conditions
• Discrete steering methods (e.g., latches, magnets) are difficult to reverse; tip-internal devices can only steer the tip; external magnets only work in small-scale environments
• While pneumatic actuators provide reversible continuous curvature control, the quantitative impact of manufacturing methods on free-space steerability remains unstudied

Research Motivation

This study fills this gap by experimentally characterizing how vine robot steerability depends on design and control parameters, focusing on robots using series pouch motor pneumatic actuators for steering.

Core Contributions

1. Systematic Analysis: First systematic analysis of how tip load, chamber pressure, length, and diameter affect vine robot steerability in self-supporting 3D conditions
2. Manufacturing Method Comparison: Comparative study of exterior actuators versus integrated actuators designs, investigating how actuator manufacturing and pressure ratio affect steerability under ground-supported planar conditions
3. Design Principle Validation: Through practical task demonstrations, proving that robots designed using these principles outperform those using ad hoc parameters
4. Generalizable Guidance: Provides generalizable design and control principles for pneumatically-steered vine robots, with reference value for other steering methods

Methodology Details

Task Definition

Steerability: The ability of vine robots to produce controlled curvature in free space—critical for maneuvering in complex environments and ensuring field performance.

This study quantifies steerability through two classes of experiments:

• Input Parameters: Tip load, chamber pressure, length, diameter, manufacturing method, actuator pressure
• Output Metrics: Characteristic length, horizontal range, vertical range, curvature

Experimental Architecture

Experiment 1: Self-Supporting 3D Steerability Experiment

Robot Manufacturing:

• Material: 30 denier heat-sealed TPU-coated nylon fabric
• Structure: Main chamber + 3 external pouch actuators uniformly distributed circumferentially
• External pouch design: Actuators individually heat-sealed then taped to the outside of the main chamber

Experimental Setup (Figure 2):

• Robot cantilevered horizontally on table with support only at base
• Motion capture LED markers track 3D tip position
• Robot sequentially visits 6 via points:
  • Individual actuator activation to maximum pressure (10.34 kPa)
  • Pairwise activation of two actuators to maximum pressure
• 3 trials per parameter set

Evaluation Metrics:

1. Characteristic Length: Square root of the convex hull area of tip position projection on x-y plane $L_c = \sqrt{\frac{1}{2}\left|\sum_{k=1}^{m}(x_k y_{k+1} - x_{k+1} y_k)\right|}$
2. Horizontal Range: Difference between maximum and minimum x-coordinates
3. Vertical Range: Difference between maximum and minimum y-coordinates

Variable Control Experiments:

1. Tip Load: 0g, 20g, 50g, 100g, 150g, 200g (fixed pressure 6.89 kPa, length 0.61m, diameter 8.1cm)
2. Chamber Pressure: 2.07-10.34 kPa (fixed load 50g, length 0.61m, diameter 8.1cm)
3. Length: 0.30m, 0.46m, 0.61m, 0.76m (fixed load 50g, pressure 6.89 kPa, diameter 8.1cm)
4. Diameter: 3.2cm, 6.5cm, 9.7cm, 12.9cm (fixed load 50g, pressure 6.89 kPa, length 0.61m)

Experiment 2: Ground-Supported Planar Steerability Experiment

Manufacturing Method Comparison (Figure 4):

1. Exterior Pouches:
   • Main chamber is single-layer TPU-coated nylon tube
   • 3 pouch actuators individually manufactured then attached to main chamber exterior
   • Actuator width is half the flattened main chamber width
2. Integrated Pouches:
   • Main chamber and actuators constructed together as dual-layer structure
   • Folded TPU-coated fabric heat-sealed to define chamber walls and internal pouch cavities
   • Actuator nominal width is 2/3 of main chamber flattened diameter

Pressure Ratio Definition: $\text{ratio} = \frac{P_{\text{pouch}}}{P_{\text{body}}}$

where $P_{\text{pouch}}$ is actuator pressure and $P_{\text{body}}$ is chamber pressure.

Experimental Setup (Figure 5):

• Robot length fixed at 1.75m, fully supported on ground
• Single actuator aligned with measurement plane
• Motion capture LED markers (25cm spacing) track planar skeleton
• Pouch and chamber pressures independently varied in 3.45-13.79 kPa range (3.45 kPa increments)
• Pressure ratio range: 0.25-4.0
• 3 trials per parameter set

Evaluation Metrics:

• Curvature calculated by least-squares circle fitting to planar tip trajectory: $\kappa = 1/r$ (r is circle radius)

Technical Innovations

1. Systematic Experimental Design: First multi-parameter systematic experiment on vine robots, using controlled variable method to study independent effects of each factor
2. 3D Free-Space Quantification: Proposes characteristic length metric to quantify 3D steerability more comprehensively than simple curvature measurement
3. Manufacturing Method Mechanical Analysis: In-depth analysis of hoop stress effects for exterior vs. integrated actuators:
   • Integrated pouches share stress with chamber walls, requiring expansion strain to overcome preload
   • Exterior pouches are mechanically decoupled, producing bending moment at lower pressure ratios
4. Pressure Ratio Normalization: Introduces pressure ratio concept providing consistent framework across designs to examine how turning force transmits through structure

Experimental Setup

Experimental Parameters

• Material: 30 denier TPU-coated tear-resistant nylon fabric
• Pressure Control: 4 closed-loop pressure regulators (Proportion-Air QB3)
• Motion Capture: PhaseSpace Impulse X2E system
• Maximum Actuator Pressure: 10.34 kPa (exterior pouch experiments)
• Pressure Range: 2.07-13.79 kPa

Experimental Procedure

1. Inflate to specified parameters before each trial
2. Record data after robot motion ceases
3. Fully deflate and reset between trials
4. Repeat each parameter set 3 times

Experimental Results

Self-Supporting 3D Steerability Results (Figure 3)

1. Tip Load Effect (Figure 3a)

• Trend: Characteristic length decreases significantly with increasing tip load
• Critical Threshold: Particularly sharp decline at 100g and above
• Horizontal and Vertical Range: Both decrease sharply with load increase
• Numerical Examples:
  • 0g load: characteristic length ~0.25m
  • 200g load: characteristic length drops to ~0.05m (80% decrease)

2. Chamber Pressure Effect (Figure 3b)

• Nonlinear Relationship: Characteristic length peaks at moderate pressure (5.52 kPa)
• Low Pressure: Insufficient robot support
• High Pressure: Robot becomes overly rigid, actuators difficult to move
• Horizontal Range: Generally decreases with increasing pressure
• Vertical Range: Irregular variation with high variance (increased rigidity prevents deflection but also impedes motion)

3. Length Effect (Figure 3c)

• Characteristic Length: Increases with robot length
• Horizontal Range: Notably increases (longer robot reaches farther with same curvature)
• Vertical Range: Decreases (longer robot heavier, harder to counteract gravity)
• Numerical Examples:
  • 0.30m: characteristic length ~0.12m
  • 0.76m: characteristic length ~0.22m (83% increase)

4. Diameter Effect (Figure 3d)

• Key Finding: Diameter has minimal impact on steerability
• Critical Threshold: 3.2cm robot collapses; other diameters show similar performance
• Design Implication: Diameter should be large enough to prevent collapse, but specific value has minimal impact

Ground-Supported Planar Steerability Results (Figure 6)

Exterior Pouch Actuators

• Low Pressure Ratio Behavior: Begin bending at lower pressure ratios (<1)
• High Pressure Ratio Behavior: Curvature saturates at high pressure ratios
• Maximum Curvature: ~0.8 m⁻¹ (at pressure ratio 4.0)

Integrated Pouch Actuators

• Activation Threshold: Begin bending only when pressure ratio exceeds 1
• Linear Growth: Curvature approximately linear with pressure ratio
• Maximum Curvature: ~1.4 m⁻¹ (at pressure ratio 4.0), 75% higher than exterior pouches

Mechanical Mechanism Explanation

• Integrated Pouches: Share stressed chamber walls; expansion must overcome hoop preload, delaying bending but ultimately achieving higher curvature
• Exterior Pouches: Mechanically decoupled, producing bending moment at low ratios, but chamber walls still resist planar shortening, causing early saturation

Hysteresis Phenomenon (Figure 7)

• Observation: Both manufacturing methods show significant residual deformation after actuator deflation
• Cause: Material compliance and interlayer friction cause permanent set
• Impact: Reduces repeatability and control precision, limiting effective 3D steerability

Task Demonstration Results (Figure 8, Table I)

Design 1: Baseline (Ad Hoc Parameters)

• Parameters: Exterior pouch, 150g load, pressure ratio 1.0
• Task 1 (Lift Height): 2cm
• Task 2 (Planar Bending): 22°

Design 2: Optimized Exterior Pouch

• Parameters: Exterior pouch, 100g load, pressure ratio 2.0 (lift)/4.0 (bending)
• Task 1: 12cm (6× improvement)
• Task 2: 68° (3× improvement)

Design 3: Optimized Integrated Pouch

• Parameters: Integrated pouch, 100g load, pressure ratio 2.0 (lift)/4.0 (bending)
• Task 1: 2cm (same as baseline)
• Task 2: 76° (best planar steering)

Key Findings:

• Reducing load and increasing pressure ratio significantly improves performance
• Exterior pouches better suited for lift-requiring tasks
• Integrated pouches perform best for ground-supported planar steering

Design and Control Recommendations

1. Diameter

• Principle: Select diameter large enough to prevent collapse under expected operating load, pressure, and length
• Impact: Minimal effect on steerability; selection primarily based on environmental requirements

2. Manufacturing Method

• Exterior Pouch Applications: Tasks requiring high pressure ratio operation (typically due to high chamber pressure needs for growth or collapse prevention)
• Integrated Pouch Applications: Tasks operable at high pressure ratio (robot supported and/or steering without active growth)

3. Length

• Principle: Maximize free length without causing collapse
• Operational Strategy: Operators can select paths, e.g., choosing earlier-encountered apertures to preserve longer length for subsequent steering

4. Tip Load

• Principle: Keep total sensor payload weight as light as possible
• Strategy: Make certain sensors detachable, assisting robot during difficult maneuvers without all sensors

5. Operating Pressure and Pressure Ratio

• Ground-Supported Planar Tasks: Maximize actuator-to-chamber pressure ratio
• Self-Supporting 3D Tasks: Select moderate chamber pressure, balancing support capability with steering flexibility
• Constraints:
  • Chamber pressure must exceed requirements for desired growth speed
  • Both chamber and actuator pressures must remain below burst pressure
• Future Directions: Improve manufacturing for higher pressure tolerance; develop control strategies coordinating growth and steering (e.g., alternating "growth phases" and "steering phases")

Related Work

Continuum Robot Research

• Actuator placement and method strongly influence deformation and workspace
• This study focuses on steering mechanism choices specific to vine robots

Vine Robot Steering Methods

1. Discrete Methods:
   • Latches, local magnets: Provide sharp curvature at predetermined locations but difficult to reverse
   • Internal tip devices: Only enable high-curvature tip steering, limited utility for free-space navigation
   • External magnets: Only applicable in small-scale environments
2. Pneumatic Actuators:
   • Series pneumatic artificial muscles (sPAMs)
   • Fabric pneumatic artificial muscles (fPAMs)
   • Cylindrical pneumatic artificial muscles
   • Series Pouch Motors: Easily manufactured via pulse heat-sealing or hot-pressing

Existing Research Gaps

• Prior research focused on unloaded curvature, growth speed, turning force
• This Study Addresses: Systematic comparison of design and control parameters for series pouch motor-driven robots in free-space 3D mobility tasks

Conclusions and Discussion

Main Conclusions

1. Tip Load: Steerability decreases significantly with load increase; 100g is critical threshold
2. Chamber Pressure: Optimal moderate pressure (~5.52 kPa) balances support and flexibility
3. Length: Longer robots achieve better steerability but must avoid excessive length causing collapse
4. Diameter: Minimal impact on steerability if sufficiently large to prevent collapse
5. Manufacturing Method:
   • Exterior pouches: Low pressure ratio activation, high pressure ratio saturation, suited for lift tasks
   • Integrated pouches: High pressure ratio activation, achieves higher maximum curvature, suited for ground steering
6. Pressure Ratio: Should be maximized or optimized based on task type

Limitations

1. Hysteresis Problem:
   • Both manufacturing methods exhibit significant residual deformation
   • Reduces repeatability and control precision
   • Requires manual intervention for reset
2. Pressure Constraints:
   • Current manufacturing limits pressure tolerance
   • Prevents simultaneous achievement of high growth speed and high curvature
   • May require switching between growth and steering
3. Experimental Conditions:
   • Self-supporting experiments only test horizontal cantilever configuration
   • Ground-supported experiments simplify gravity effects
   • Real-world environments more complex
4. Sensor Deficiency:
   • Lacks proprioceptive sensing of robot shape
   • Operators cannot real-time assess residual curvature

Future Directions

1. Low-Hysteresis Design:
   • Explore robot body designs reducing material compliance and interlayer friction
   • Develop improved heat-sealing techniques
2. Proprioceptive Sensing:
   • Integrate sensors detecting robot shape
   • Provide real-time shape feedback to operators
   • Compensate for residual deformation
3. High-Pressure Actuators:
   • Develop actuator designs tolerating higher pressures
   • Maintain compatibility with vine robot growth
4. Time-Varying Control Strategies:
   • Develop time-varying actuator-to-chamber pressure ratio control
   • Selectively enable faster growth or greater steering per operator needs
   • Coordinate "growth phases" and "steering phases"
5. Extended Applications:
   • Generalize results to other pneumatically-steered vine robots
   • Provide reference for alternative steering methods

In-Depth Evaluation

Strengths

1. Strong Systematicity:
   • First multi-parameter systematic experimental study of vine robots
   • Rigorous controlled variable methodology with independent parameter variation
   • Comprehensive experimental design covering both 3D self-supporting and 2D ground-supported scenarios
2. High Practical Value:
   • Directly addresses real-world field application challenges (load, pressure selection)
   • Provides explicit design and control recommendations
   • Task demonstrations validate guideline effectiveness
3. Deep Physical Insight:
   • Clear mechanical explanation of manufacturing method differences (hoop stress effects)
   • Pressure ratio normalization concept facilitates cross-design comparison
   • Identifies nonlinear relationships (e.g., non-monotonic pressure effects)
4. Experimental Design Innovation:
   • Characteristic length metric effectively quantifies 3D steerability
   • Motion capture system precisely tracks 3D trajectories
   • Good repeatability (3 trials per parameter set)
5. Clear Writing:
   • Rich, information-dense figures
   • Detailed methodology enabling reproducibility
   • Clear, actionable conclusions

Weaknesses

1. Missing Theoretical Models:
   • Primarily relies on experimental observation, lacks predictive mathematical models
   • No quantitative relationships established between parameters and steerability
   • Difficult to extrapolate to untested parameter combinations
2. Unresolved Hysteresis Problem:
   • Identifies residual deformation but offers no solutions
   • Severely impacts control precision in practical applications
   • Requires further research understanding and mitigating hysteresis
3. Limited Experimental Scope:
   • Tests only one actuator type (series pouch motor)
   • Self-supporting experiments only one posture (horizontal cantilever)
   • No dynamic motion or rapid maneuvering tests
4. Simplified Environmental Factors:
   • Experiments conducted under ideal laboratory conditions
   • Ignores obstacles, friction, uneven surfaces in real environments
   • Field deployment may encounter greater complexity
5. Sensing and Control Limitations:
   • No closed-loop control experiments
   • No integrated shape sensing feedback
   • Control strategies remain open-loop pressure settings
6. Limited Statistical Analysis:
   • No significance testing performed
   • Error bars show only standard deviation, not confidence intervals
   • Small sample size (3 trials per parameter set)

Impact

1. Academic Contribution:
   • Provides systematic experimental data for soft robotics field
   • First quantification of multiple design parameters' effects on vine robot performance
   • Establishes experimental foundation for subsequent theoretical modeling
2. Practical Value:
   • Directly guides search and rescue robot design
   • Reduces trial-and-error costs and development time
   • Improves field deployment success rates
3. Reproducibility:
   • Detailed methodology description
   • Complete material and equipment information
   • Code and data availability not explicitly stated (potential weakness)
4. Field Advancement:
   • Advances vine robots from proof-of-concept to engineering applications
   • Stimulates follow-up research (theoretical modeling, new actuator designs)
   • May influence other soft robot designs

Applicable Scenarios

1. Urban Search and Rescue:
   • Confined space exploration
   • Carrying lightweight sensors (<100g)
   • Requiring 3D maneuvering capability
2. Pipeline Inspection:
   • Ground-supported conditions
   • Integrated pouch design enables high-curvature steering
   • Long-distance extension
3. Archaeological Exploration:
   • Gentle contact needed
   • Possible lift capability requirements
   • Exterior pouch design more suitable
4. Medical Applications:
   • Minimally invasive surgery navigation
   • Precise control required
   • Hysteresis problem needs resolution
5. Unsuitable Scenarios:
   • High-speed motion tasks
   • Heavy payload requirements (>200g)
   • Demanding high repeatability precision tasks

Key References

1 Blumenschein et al., 2020 - Comprehensive review of vine robot design, modeling, control, and applications

4 Coad et al., 2020 - Vine robot design, teleoperation, and field deployment

6 Hawkes et al., 2017 - Soft robot navigating environments through growth

8 Ataka et al., 2020 - Model-based posture control of variable-stiffness inflatable eversion robots

9 Kübler et al., 2024 - Comparison of pneumatic actuators for soft growing vine robots

17 Greer et al., 2017 - Series pneumatic artificial muscles (sPAMs) and applications in soft continuum robots

Overall Assessment: This is a high-quality experimental robotics paper that systematically investigates factors affecting vine robot steerability, providing valuable design guidance for the field. The paper's main strengths lie in comprehensive experimental design and practical conclusions; main weaknesses are absence of theoretical models and lack of hysteresis solutions. This work significantly advances vine robots from laboratory research toward practical field applications.