Transporter Networks: Rearranging the Visual World for Robotic Manipulation

TL;DR

Transporter Networks are a novel architecture for robotic manipulation that rearranges visual features to infer spatial displacements, enabling efficient learning and generalization across diverse manipulation tasks without relying on object assumptions.

Contribution

The paper introduces the Transporter Network, a new model architecture that efficiently learns visual manipulation policies by rearranging deep features, outperforming existing methods in sample efficiency and generalization.

Findings

Learns faster than end-to-end baselines.

Generalizes to multi-step and 6DoF tasks.

Effective in real-world robotic experiments.

Abstract

Robotic manipulation can be formulated as inducing a sequence of spatial displacements: where the space being moved can encompass an object, part of an object, or end effector. In this work, we propose the Transporter Network, a simple model architecture that rearranges deep features to infer spatial displacements from visual input - which can parameterize robot actions. It makes no assumptions of objectness (e.g. canonical poses, models, or keypoints), it exploits spatial symmetries, and is orders of magnitude more sample efficient than our benchmarked alternatives in learning vision-based manipulation tasks: from stacking a pyramid of blocks, to assembling kits with unseen objects; from manipulating deformable ropes, to pushing piles of small objects with closed-loop feedback. Our method can represent complex multi-modal policy distributions and generalizes to multi-step sequential tasks, as well as 6DoF pick-and-place. Experiments on 10 simulated tasks show that it learns faster and generalizes better than a variety of end-to-end baselines, including policies that use ground-truth object poses. We validate our methods with hardware in the real world. Experiment videos and code are available at https://transporternets.github.io