Modular compliant robotics

Lego-Like
Stiffness Configuration
of Planar Compliant Modules
for Task-Specific Flexible Interfaces

Explore the design

Manual deflection · Gripper inset: 3× Watch full-frame clip ↗

Siyue Yao1,2, Xiaochi Xie2, Shixuan Zhao2, Yutong Li1,
Hao Li2, Mark R. Cutkosky2, Genliang Chen3,4,*

1,3,4 Shanghai Jiao Tong University  ·  2 Stanford University

Submitted to ICRA 2027 · Under review

The idea

Build the stiffness
a task needs.

A flexible wrist should bend in the directions that help a task, while staying stiff where support is needed. We configure that behavior by stacking a small set of planar compliant modules.

Module geometry, beam width, plate thickness, and orientation determine the stiffness of the assembly. A unified model and two-stage optimization choose a composition for a desired stiffness profile. The same building blocks produce distinct contact responses without redesigning the complete mechanism.

01 / From planar parts to a flexible interface

Three geometries.
Many stiffness profiles.

Curved, spiral, and wavy beams give the three module types complementary stiffness characteristics. Adjusting their dimensions changes the balance between translation and rotation; changing their orientation swaps directional stiffness.

Click a module to flip it. Open all three to compare. Swipe to see the other modules.

Illustrative profiles redrawn from Figure 2B on a shared scale. Rotational terms: k*ri = 5kri/R². Original figure ↗

From a target to an assembly

A genetic algorithm selects module configurations, then sequential quadratic programming refines the continuous dimensions. The optimization can target specific stiffness values or relative preferences between components.

01

Specify

Target stiffness and geometric constraints

02

Compose

Module types, orientations, and dimensions

03

Stack

A flexible interface with the chosen stiffness

See the inverse design results
Optimization flowchart and radar plots comparing optimized stiffness with different task requirements.
The paper evaluates prescribed stiffness values and qualitative stiffness preferences.

02 / Watch the contact response

Changing stiffness changes
what happens on contact.

A wrist with low rotational stiffness can reorient around an obstacle, but exhibits noticeable oscillation. A wrist with higher rotational stiffness resists reorientation and eventually tips the obstacle. The optimized composition balances compliance with the desired dynamic response.

Left: optimized composite wrist, real time. Right: a single module with low rotational stiffness at 2 mm/s, shown first at 10× and then in real time to reveal oscillation.

Compliant interaction at 1 m/s

The optimized wrist targets a rotational stiffness of 5 N·m/rad and a translational stiffness of 6,000 N/m. It combines a Type-I module with two mirrored Type-II modules.

In the high-speed experiment, the robot moves back and forth at 1 m/s with a maximum imposed vertical offset of 10 mm. The wrist accommodates the obstacle on each contact. Approximately 15° of angular compliance is demonstrated in the tested condition.

Optimized modular wrist during high-speed interaction, with close-ups of the compliant stack. Main view: real time.
See the optimized wrist assembly
Optimized stiffness radar plot, modular composition, and wrist interacting with an obstacle.
Target and optimized stiffness, the modular composition, and the corresponding wrist experiment.

03 / Check the model

Measured across
six stiffness components.

A fabricated Type-I module is loaded separately in six configurations. Each test repeats 20 times. Near the unloaded configuration, the measured force–displacement curves are fitted to identify the module stiffness.

Stiffness characterization under different loading directions. Playback: 20×.

After calibrating the material's elastic modulus, the model's deviation from the measured stiffness is below 6.5% for all six components, with an average deviation below 5%. Linear fits near the unloaded configuration achieve R² ≥ 0.99.

Force-displacement measurements, simulated and experimental stiffness values, and relative errors below 6.5 percent across six stiffness components.
Experimental characterization and comparison with the stiffness model. The model uses calibrated material properties.
Scope of the current model

The framework models stiffness near the unloaded configuration and targets a diagonal stiffness matrix. The underlying planar structures still couple some stiffness components, so arbitrary profiles are not always attainable. Stack thickness, module offsets, and geometric interference can also affect the assembled behavior.

Paper & authors

A modular approach
to task-specific compliance.

Lego-Like Stiffness Configuration of Planar Compliant Modules for Task-Specific Flexible Interfaces

Siyue Yao, Xiaochi Xie, Shixuan Zhao, Yutong Li, Hao Li, Mark R. Cutkosky, and Genliang Chen.

Submitted to ICRA 2027 · Under review

arXiv:2610.07319 · v1 · October 5, 2026

Citation / BibTeX

@misc{yao2026legolikestiffnessconfigurationplanar,
      title={Lego-Like Stiffness Configuration of Planar Compliant Modules for Task-Specific Flexible Interfaces},
      author={Siyue Yao and Xiaochi Xie and Shixuan Zhao and Yutong Li and Hao Li and Mark R. Cutkosky and Genliang Chen},
      year={2026},
      eprint={2610.07319},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2610.07319},
}
Full affiliations
  1. State Key Laboratory of Mechanical Systems and Vibration, Shanghai Jiao Tong University, Shanghai, China
  2. Department of Mechanical Engineering, Stanford University, USA
  3. Shanghai Key Laboratory of Intelligent Robotics, Shanghai Jiao Tong University, Shanghai, China
  4. META Robotics Institute, Shanghai Jiao Tong University, Shanghai, China

* Corresponding author: Genliang Chen

The complete experiment video

See the full sequence.

Stiffness verification, manual compliance, and side-by-side contact experiments in the original 49-second video.