TeCoBot

Play with TeCoBots

for visualization purposes

TeCoBots

Modular tensegrity continuum robots that bend, collapse, and team up by docking.

Tensegrity Continuum Robots Enable Task-Adaptive Morphologies for Cooperative Behaviors
Nature Machine Intelligence (2026)

M. H. Saikot, S. Spiegel, S. A. Kariyawasam, A. Stefka, J. Chrisler, J. Zhao

Adaptive Robotics Lab, Colorado State University

scroll to read more ↓

Robot

Carbon-fibre rod
Spacer disk
Modules
Body length
Overall length

About this page

Drag to orbit, wheel to zoom. Key names the parts on the robot itself. Frame resizes the cuboid and slides each docking station along the edge it is bolted to. Cables drives a robot by its three cable displacements: pull all three together and the body shortens dead straight, pull them unevenly and the difference is what bends it.

Where a configuration does not offer the control of one of its robots, that robot’s pose is held fixed so a dock cannot come apart under the sliders — the hardware has no such restriction, and can be driven in ways this page does not reproduce. Nothing here checks for collision. Every dimension comes from the MM block at the top of the source; correct a number there and the whole model follows.

In under a minute

Three short reels of TeCoBots at work, alone and together (sped up).

Abstract

Robots that can change their morphologies and behaviors for different tasks and environments hold great promise for adaptable, multifunctional systems. Modular reconfigurable robots (MRRs) can achieve such functionalities by docking and rearranging individual units, but most rely on rigid modules that lack structural compliance, resulting in limited capabilities. Continuum robots offer compliance through flexible backbones, yet they cannot self-reconfigure into task-adaptive multi-robot configurations. Here, we introduce an MRR that unifies the advantages of both architectures by combining a tensegrity-based compliant body with claw-based connection mechanisms. Each robot can manipulate and locomote independently, and multiple robots can self-reconfigure into different morphologies (e.g., chains, loops, branches) for cooperative manipulation and locomotion. We demonstrate the robots’ capability across diverse tasks and environments, including coordinated object manipulation and transport, multimodal locomotion, and loco-manipulation in real-world scenarios. These results lay a foundation for adaptable and multifunctional robotic collectives, with broad potential applications in manufacturing, space exploration, and search-and-rescue operations.

What’s new

Three ideas behind TeCoBot. Open any of them for the details.

Six TeCoBots docked into an elephant: a spine, four legs and a trunk.

Imagine

What it could become

Six robots as an elephant: one spine, four legs and a trunk, joined by the claws.

Try it in Play

One robot

What a single TeCoBot can do: handle objects (manipulate), move six different ways, and roll across various terrain (sped up).

Manipulation

Shortening the bodyExtends to grab grapes overhead, then collapses and bends its body to drop them in the jar.
BendingBends toward grapes on either side, then bends again to place them in the jar.
Load bearingHanging from one gripper, the robot holds and lifts a full 2.5 L water bottle with the other.

Six ways to move

Axial extension–compressionClaws anchor in turn while the body shortens and lengthens.
InchwormThe body arches and slides forward on directional-friction claws.
Crawling with clawsOne claw opens and closes against the ground to drag the body along.
RollingCables pull in sequence, shifting the centre of mass so the robot tips over and rolls. The fastest gait.
TurningOne cable bends the body toward the turn between claw anchors.
Pipe climbingAxial extension–compression inside a pipe carries it straight up.

Eight terrains tested

Rolling works best: it keeps going over loose, uneven ground like rocks and hay, and is fastest on smooth wood, at 30 mm/s.

Sidewalk
Grass
Gravel
Hay
Rocks
Soil
Smooth wood
Low-pile carpet

Cooperative manipulation

Three robots on a frame can rearrange themselves into many layouts, each suited to a different job (sped up).

M1: lifting a ballThree hanging robots bend inward together to cooperatively lift a 200 mm ball.
M2: test tubePick up a test tube, remove the cork and pour liquid into a beaker.
M3: tissueTwo ground robots hold the roll while the vertical robot unrolls and tears.
M3: open containerGround robots hold the container steady while the vertical robot removes the lid and retrieves the object inside.
M4: extended reachThree in series reach a bottle beyond any one robot’s reach.
M5: orienting a blockThree robots lying on the ground work together, pushing and rolling a cube to turn it to a new face.

Cooperative locomotion

On the ground, robots link into chains, loops and branches to move in ways one robot cannot (sped up).

O-: carrying a large objectThe loop encloses the object while the tail inches it forward.
H: climbing a ductParallel robots anchor in turn while the cross-member pushes and pulls.
O: rolling around a pipeThree robots closed into a loop roll along a large cylinder.
Y: object transportTwo branch robots hold a box while the rear robot drives.
Square: four robotsAn enclosed square moves forward as one body.
Chain: rolling over a logOutdoors, a chain rolls over obstacles a single robot could not.

Real-world tasks

The robots climb, hand over, and carry objects, changing configurations between steps (sped up).

Climbing onto the tableTwo robots push the third up onto the tabletop.
Handover between levelsOne robot passes scrap paper from the tabletop to another robot standing on the floor.
Forming a chain across levelsA floor robot docks to the one on the table.
Into the binThe chain carries the scrap paper over and drops it in.
Cabinet organizationA part is handed over and placed in the drawer.
Sample collection outdoorsTwo robots roll forward, then one stands up to pluck a flower and place it in a beaker while remaining connected.

Videos

The full videos, with the playback speed noted for each.

Cooperative manipulationMultiple robots work together to grasp, move and handle objects, changing how they are arranged to suit each task.

Go further

Citation

@article{saikot2026tecobot,
  title   = {Tensegrity continuum robots enable task-adaptive
             morphologies for cooperative behaviours},
  author  = {Saikot, Mahmud Hasan and Spiegel, S. and Kariyawasam, S. A.
             and Stefka, A. and Chrisler, J. and Zhao, Jianguo},
  journal = {Nature Machine Intelligence},
  year    = {2026},
  doi     = {10.1038/s42256-026-01335-4},
  url     = {https://www.nature.com/articles/s42256-026-01335-4}
}