# Learning to Design and Construct Bridge without Blueprint

arXiv: https://arxiv.org/abs/2108.02439 · HTML: https://mlsanity.com/paper/2108.02439/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 35 references that are in the map.
- First posted on arXiv 5 Aug 2021.

## Builds on (most cited first, 20 of 20)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Solving Rubik's Cube with a Robot Hand](https://mlsanity.com/paper/1910.07113/) (2019)
- [Open3D: A Modern Library for 3D Data Processing](https://mlsanity.com/paper/1801.09847/) (2018)
- [Interaction Networks for Learning about Objects, Relations and Physics](https://mlsanity.com/paper/1612.00222/) (2016)
- [Automatic Goal Generation for Reinforcement Learning Agents](https://mlsanity.com/paper/1705.06366/) (2017)
- [Reverse Curriculum Generation for Reinforcement Learning](https://mlsanity.com/paper/1707.05300/) (2017)
- [Active Domain Randomization](https://mlsanity.com/paper/1904.04762/) (2019)
- [Intrinsic Motivation and Automatic Curricula via Asymmetric Self-Play](https://mlsanity.com/paper/1703.05407/) (2017)
- [Phasic Policy Gradient](https://mlsanity.com/paper/2009.04416/) (2020)
- [MCP: Learning Composable Hierarchical Control with Multiplicative Compositional Policies](https://mlsanity.com/paper/1905.09808/) (2019)
- [Hierarchical Planning for Long-Horizon Manipulation with Geometric and Symbolic Scene Graphs](https://mlsanity.com/paper/2012.07277/) (2020)
- [Composable Deep Reinforcement Learning for Robotic Manipulation](https://mlsanity.com/paper/1803.06773/) (2018)
- [Towards Practical Multi-Object Manipulation using Relational Reinforcement Learning](https://mlsanity.com/paper/1912.11032/) (2019)
- [Form2Fit: Learning Shape Priors for Generalizable Assembly from Disassembly](https://mlsanity.com/paper/1910.13675/) (2019)
- [Online 3D Bin Packing with Constrained Deep Reinforcement Learning](https://mlsanity.com/paper/2006.14978/) (2020)
- [Reinforcement Learning on Variable Impedance Controller for High-Precision Robotic Assembly](https://mlsanity.com/paper/1903.01066/) (2019)
- [Revisiting Design Choices in Proximal Policy Optimization](https://mlsanity.com/paper/2009.10897/) (2020)
- [Solving Compositional Reinforcement Learning Problems via Task Reduction](https://mlsanity.com/paper/2103.07607/) (2021)
- [Refactoring Policy for Compositional Generalizability using Self-Supervised Object Proposals](https://mlsanity.com/paper/2011.00971/) (2020)

## Built on by (most cited first, 6 of 6)

- [VLMgineer: Vision Language Models as Robotic Toolsmiths](https://mlsanity.com/paper/2507.12644/) (2025)
- [Learning to Design and Use Tools for Robotic Manipulation](https://mlsanity.com/paper/2311.00754/) (2023)
- [Budget-Aware Sequential Brick Assembly with Efficient Constraint Satisfaction](https://mlsanity.com/paper/2210.01021/) (2022)
- [RobotSmith: Generative Robotic Tool Design for Acquisition of Complex Manipulation Skills](https://mlsanity.com/paper/2506.14763/) (2025)
- [Learning Design and Construction with Varying-Sized Materials via Prioritized Memory Resets](https://mlsanity.com/paper/2204.05509/) (2022)
- [Learning to Build: Autonomous Robotic Assembly of Stable Structures Without Predefined Plans](https://mlsanity.com/paper/2602.23934/) (2026)

## Abstract (arXiv)

Autonomous assembly has been a desired functionality of many intelligent robot systems. We study a new challenging assembly task, designing and constructing a bridge without a blueprint. In this task, the robot needs to first design a feasible bridge architecture for arbitrarily wide cliffs and then manipulate the blocks reliably to construct a stable bridge according to the proposed design. In this paper, we propose a bi-level approach to tackle this task. At the high level, the system learns a bridge blueprint policy in a physical simulator using deep reinforcement learning and curriculum learning. A policy is represented as an attention-based neural network with object-centric input, which enables generalization to different numbers of blocks and cliff widths. For low-level control, we implement a motion-planning-based policy for real-robot motion control, which can be directly combined with a trained blueprint policy for real-world bridge construction without tuning. In our field study, our bi-level robot system demonstrates the capability of manipulating blocks to construct a diverse set of bridges with different architectures.