Learning to Design and Construct Bridge without Blueprint

08/05/2021
by   Yunfei Li, et al.
0

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.

READ FULL TEXT

page 1

page 6

research
04/12/2022

Learning Design and Construction with Varying-Sized Materials via Prioritized Memory Resets

Can a robot autonomously learn to design and construct a bridge from var...
research
03/04/2022

Cooperative Task and Motion Planning for Multi-Arm Assembly Systems

Multi-robot assembly systems are becoming increasingly appealing in manu...
research
03/08/2022

Graph-based Reinforcement Learning meets Mixed Integer Programs: An application to 3D robot assembly discovery

Robot assembly discovery is a challenging problem that lives at the inte...
research
08/01/2022

Hierarchical Reinforcement Learning for Precise Soccer Shooting Skills using a Quadrupedal Robot

We address the problem of enabling quadrupedal robots to perform precise...
research
11/20/2022

Multi-Arm Bin-Picking in Real-Time: A Combined Task and Motion Planning Approach

Automated bin-picking is a prerequisite for fully automated manufacturin...
research
11/10/2017

FunTAL: Reasonably Mixing a Functional Language with Assembly

We present FunTAL, the first multi-language system to formalize safe int...
research
09/26/2018

Pay attention! - Robustifying a Deep Visuomotor Policy through Task-Focused Attention

Several recent projects demonstrated the promise of end-to-end learned d...

Please sign up or login with your details

Forgot password? Click here to reset