Pipes, Problems, and Prior Limits
Pipeline networks form the hidden circulatory system of modern industry, carrying oil, gas, and water across vast distances and through buried, often inaccessible terrain. When something goes wrong inside these networks, the damage is rarely spotted until a leak has already caused costly harm.
Automated inspection robots offer a way to look inside without exposing humans to risk, but the internal geometry of real pipeline systems is highly hostile to conventional machines. Sharp turns, branching junctions, and diameters that shift from one segment to the next are all routine features.
Existing robot designs struggle with at least one of these challenges. Rigid-bodied robots built around electromagnetic motors and gear drives perform well in large, straight pipes but cannot negotiate tight bends. Soft robots constructed from compliant materials fare better in confined spaces, but most are tethered, limiting how far they can travel.
Tethered systems are also restricted to passive turning, as they follow a curve rather than actively choosing a direction. Magnetically driven robots require an external field source, which rules them out for industrial-scale pipelines. Vine-style robots that grow from the tip extend well but struggle to retract. Very few soft robots have been successfully deployed for inspection in real-world pipeline networks at sub-decimeter scales.
Earthworms, Origami, and Honeycomb
SPPIRO borrows its movement logic from earthworms. A central propulsion section, 68.5 mm in diameter and 120 mm in length, contains three pneumatic Kresling origami actuators (KOAs), arranged 120 degrees apart. Kresling origami structures have an unusual mechanical property, meaning that as they contract, they twist by a predictable amount.
Each KOA is fitted with a magnetic encoder that measures the rotational angle. Because the relationship between twist angle and contraction length is deterministic, the robot can calculate the current length of each actuator without additional sensors. Feeding these three lengths into a piecewise constant-curvature (PCC) model allows the robot to reconstruct both its own shape and the pipe's geometry in real time.
At each end of the propulsion section sits an anchor module. Each module uses a brushed direct current (DC) motor to drive a six-bar compound linkage that expands radially to more than twice the robot's torso diameter, allowing SPPIRO to brace against pipe walls ranging from 140 mm to 200 mm in diameter.
The contact feet are three-dimensional (3D)-printed from 50A-hardness resin in a honeycomb lattice, then coated with Ecoflex silicone. The lattice conforms passively to the pipe wall under radial load while remaining stiff enough axially to transfer the push-pull forces of locomotion.
Each honeycomb wheel also embeds a liquid-metal capacitive sensor, allowing the two conductive layers to compress together when the wheel presses against the wall, causing capacitance to shift by up to 25% and providing the operator with a clear contact signal even in an opaque environment. All electronics are integrated onto two custom printed circuit boards (PCBs), enabling continuous untethered operation for 26 min.
Speed, Turns, and Surface Grime
Testing across pipes ranging from 145 mm to 195 mm in diameter showed SPPIRO reaching a peak vertical speed of 9.5 mm/s in the tightest pipe, with speeds of around 7.0-8.5 mm/s in larger diameters. Horizontal speeds were consistently 1-2 mm/s lower than vertical, a gap the authors attribute to the robot's body sagging away from the pipe axis when gravity acts sideways, which drags the non-anchoring foot along the wall and adds friction.
Download the PDF of this page here
Despite being fully untethered and pneumatically driven, SPPIRO's speed measured in body lengths per second is approximately ten times higher than that of comparable untethered soft robots. In complex terrain tests, SPPIRO navigated a 90-degree sharp turn in a 145 mm pipe in under 94 s, and completed a full T-junction traverse in 166 s.
The robot also traversed a 1,050 mm multi-segment pipe spanning diameters of 145 mm, 175 mm, and 195 mm, reconstructing the geometry throughout. Robustness tests across four surface conditions (dry polyvinyl chloride (PVC), wet PVC, clean metal, and mud-coated rusted metal) showed average speeds of 9.5, 10.1, 8.5, and 6.9 mm/s, respectively, demonstrating that performance holds up under realistic contamination, though soiled surfaces introduced the greatest friction losses.
Pathways and Pending Problems
SPPIRO marks the first untethered soft robot to actively steer through a T-junction, a milestone that moves the technology meaningfully closer to real-world deployment. The PCC model enables pipe geometry reconstruction during motion, and liquid-metal contact sensing supports navigation in fully opaque networks.
The authors identify three key limitations: the three-foot anchor geometry can produce brief contact singularities during turning that require time-consuming self-correction; the torso sags under full contraction during horizontal travel, increasing friction; and the PCC model becomes less accurate when frequent wall contact violates its smooth-curvature assumption. Planned improvements include stiffer actuator designs, segmented multi-body configurations, and onboard autonomous navigation through simultaneous localisation and mapping (SLAM) integration.
Journal Reference
Chen, Z., Zhang, X., Wang, J., & Huang, X. (2026). SPPIRO: a Soft, Power-autonomous, Proprioceptive In-pipe Robot for adaptive inspection in complex pipeline environments. Npj Robotics. DOI:10.1038/s44182-026-00093-0, https://www.nature.com/articles/s44182-026-00093-0
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.