The task of developing robotic systems for environmental monitoring in wetlands and swamps requires operation across diverse terrestrial and aquatic conditions. Conventional fixed-geometry wheels cannot adapt their contact geometry to these varying environments. This work presents MOSS, a compact amphibious robotic platform with 3D-printed origami wheels that address this limitation.
The platform and its adaptive wheel geometry
MOSS features a compact crank mechanism that switches each wheel between contracted and expanded configurations. This transition changes the wheel's width, body diameter, and fin orientation simultaneously. The contracted configuration presents a narrower profile suited to efficient rolling on firmer ground. The expanded configuration increases contact area and terrain engagement on softer surfaces. Both configurations are produced from the same printed hardware; only the crank position differs.
Each wheel's geometry directly influences the robot's interaction with the ground. In the contracted state, the reduced width and altered fin angle lower rolling resistance on compacted soil and mud. In the expanded state, the larger contact patch and modified fin splay distribute load over a broader area, which improves traction on loose sand, wet mud, and water surfaces. The mechanism requires approximately 1.95 seconds to complete a full configuration change, allowing morphology transitions during motion when needed.
Experimental evaluation across terrain types
The authors evaluated MOSS on soil, sand, mud with varying water content, and open water. Each terrain type presents distinct challenges: soil and firm mud favor efficient rolling with minimal deformation, sand requires resistance to slip and sinkage, wet mud combines fluidity with reduced cohesion, and water demands propulsion mechanisms adapted to buoyancy and drag.
Quantified results show a clear terrain-dependent advantage for each configuration. On firmer soil and mud, the contracted wheel achieved a lower cost of transport, meaning less energy per meter traveled. On wetter mud, sand, and water, the expanded configuration outperformed the contracted state, reducing the cost of transport by up to 29 percent. Expanded wheels also reduced slip in wet mud and reduced both slip and hub sinkage in sand, two factors that limit progress on loose or fluid substrates.
Beyond static surface tests, the platform demonstrated in-motion morphing, where wheel configuration changed while the robot continued moving. It also completed terrestrial-aquatic transitions, traversing from dry land through muddy zones into open water without requiring a full stop. Outdoor trials confirmed operation across natural, deformable terrain and variable water conditions, confirming that the adaptive geometry functions outside controlled laboratory settings.
Design implications and broader potential
The results demonstrate that adaptive wheel geometry provides terrain-dependent locomotion benefits. A single compact platform can traverse substantially different terrestrial and aquatic conditions by switching between two predefined configurations. This eliminates the need for multiple specialized robots when deploying systems to heterogeneous environments like swamps, river deltas, or coastal wetlands.
For developers designing robotic systems for similar domains, the key insight is that contact geometry modulation can outperform fixed geometries across varied substrates. The crank mechanism's speed—under 2 seconds per transition—enables quasi-continuous adaptation, which is critical for maintaining locomotion when terrain conditions change gradually. The 3D-printed origami wheel architecture also suggests that rapid prototyping of custom geometries is feasible for domain-specific applications.
One practical consideration is the trade-off between configuration speed and terrain adaptation range. The MOSS system balances these by offering two distinct configurations with a transition time fast enough to keep pace with natural terrain transitions. Applications beyond environmental monitoring could include search-and-rescue deployment in flood-prone areas, agricultural field scouting across soil type variations, or aquatic-terrestrial surveillance in conservation areas.