Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Lightweight and Transferable Design for Robust LEGO Manipulation

Ruixuan Liu, Yifan Sun, Changliu Liu

TL;DR

The paper tackles the problem of fast, safe robotic Lego prototyping by introducing a hardware-software co-design that centers on a modular EOAT capable of insert-and-twist Lego manipulation. It formalizes a safe optimization of the EOAT parameters using CMA-ES, minimizing a cost $L(\theta, d_x, d_z, T, U)$ that balances speed, precision, and safety, and validates the approach on multiple industrial robots. Key contributions include the EOAT design that reduces manipulation complexity to a small parameter set, a controller-agnostic safe-learning framework, and demonstrated transferability across robot platforms, enabling sustainable rapid Lego prototyping. The work demonstrates 100% assembly success and robust disassembly performance, highlighting practical impact for automated prototyping in manufacturing and education contexts with reduced risk and manual intervention.

Abstract

Lego is a well-known platform for prototyping pixelized objects. However, robotic Lego prototyping (i.e., manipulating Lego bricks) is challenging due to the tight connections and accuracy requirements. This paper investigates safe and efficient robotic Lego manipulation. In particular, this paper reduces the complexity of the manipulation by hardware-software co-design. An end-of-arm tool (EOAT) is designed, which reduces the problem dimension and allows large industrial robots to manipulate small Lego bricks. In addition, this paper uses evolution strategy to optimize the robot motion for Lego manipulation. Experiments demonstrate that the EOAT can reliably manipulate Lego bricks and the learning framework can effectively and safely improve the manipulation performance to a 100% success rate. The co-design is deployed to multiple robots (i.e., FANUC LR-mate 200id/7L and Yaskawa GP4) to demonstrate its generalizability and transferability. In the end, we show that the proposed solution enables sustainable robotic Lego prototyping, in which the robot can repeatedly assemble and disassemble different prototypes.

A Lightweight and Transferable Design for Robust LEGO Manipulation

TL;DR

The paper tackles the problem of fast, safe robotic Lego prototyping by introducing a hardware-software co-design that centers on a modular EOAT capable of insert-and-twist Lego manipulation. It formalizes a safe optimization of the EOAT parameters using CMA-ES, minimizing a cost that balances speed, precision, and safety, and validates the approach on multiple industrial robots. Key contributions include the EOAT design that reduces manipulation complexity to a small parameter set, a controller-agnostic safe-learning framework, and demonstrated transferability across robot platforms, enabling sustainable rapid Lego prototyping. The work demonstrates 100% assembly success and robust disassembly performance, highlighting practical impact for automated prototyping in manufacturing and education contexts with reduced risk and manual intervention.

Abstract

Lego is a well-known platform for prototyping pixelized objects. However, robotic Lego prototyping (i.e., manipulating Lego bricks) is challenging due to the tight connections and accuracy requirements. This paper investigates safe and efficient robotic Lego manipulation. In particular, this paper reduces the complexity of the manipulation by hardware-software co-design. An end-of-arm tool (EOAT) is designed, which reduces the problem dimension and allows large industrial robots to manipulate small Lego bricks. In addition, this paper uses evolution strategy to optimize the robot motion for Lego manipulation. Experiments demonstrate that the EOAT can reliably manipulate Lego bricks and the learning framework can effectively and safely improve the manipulation performance to a 100% success rate. The co-design is deployed to multiple robots (i.e., FANUC LR-mate 200id/7L and Yaskawa GP4) to demonstrate its generalizability and transferability. In the end, we show that the proposed solution enables sustainable robotic Lego prototyping, in which the robot can repeatedly assemble and disassemble different prototypes.
Paper Structure (11 sections, 2 equations, 6 figures, 1 table, 1 algorithm)

This paper contains 11 sections, 2 equations, 6 figures, 1 table, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORKS
  3. END-OF-ARM TOOL DESIGN
  4. ROBOT LEARNING FOR LEGO MANIPULATION
  5. EXPERIMENTS
  6. Transferability to Different Robot Platforms
  7. Sustainable Robotic Lego Prototyping
  8. DISCUSSION
  9. Limitations
  10. Future works
  11. CONCLUSION

Figures (6)

  • Figure 1: Top: Lego manipulation requirements. Bottom: Examples of 2D and 3D Lego prototypes.
  • Figure 2: EOAT design for Lego manipulation.
  • Figure 3: Left: Experiment environment setup. Right: Lego structures for testing manipulation performance.
  • Figure 4: Safe manipulation learning costs and parameters evolution. Left: Joint JPC. Right: Cartesian JPC. Top: learning costs. Middle: disassembly parameters. Bottom: assembly parameters. Joint JPC disassembly parameters: $T=0.6$s, $\theta=11^\circ$, $d_x=0$mm, $d_z=9.6$mm. Joint JPC assembly parameters: $T=0.4$s, $\theta=12^\circ$, $d_x=8.1$mm, $d_z=0$mm. Cartesian JPC disassembly parameters: $T=6.1$s, $\theta=12^\circ$, $d_x=0$mm, $d_z=9.5$mm. Cartesian JPC assembly parameters: $T=5.9$s, $\theta=12^\circ$, $d_x=7.8$mm, $d_z=0$mm.
  • Figure 5: Transferable to different robot platforms: Yaskawa GP4 robots manipulating Lego bricks.
  • ...and 1 more figures