Learning to Design and Construct Bridge without Blueprint

TL;DR

This paper presents a bi-level system enabling robots to autonomously design and construct bridges without blueprints, using deep reinforcement learning for planning and motion planning for execution, demonstrating versatility in real-world tasks.

Contribution

The paper introduces a novel bi-level approach combining deep reinforcement learning and motion planning for autonomous bridge construction without blueprints, with generalization to various configurations.

Findings

Successful real-world bridge construction with diverse architectures

Generalization of the blueprint policy to different block numbers and cliff widths

Effective integration of learned planning and motion control policies

Abstract

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.