RiEMann: Near Real-Time SE(3)-Equivariant Robot Manipulation without Point Cloud Segmentation

TL;DR

RiEMann is a real-time SE(3)-equivariant imitation learning framework for robot manipulation from scene point clouds, capable of generalizing to unseen objects and transformations without segmentation, and outperforming baselines in success rate and pose accuracy.

Contribution

RiEMann introduces a novel end-to-end framework that predicts target poses directly from point clouds, eliminating the need for segmentation and enabling near real-time manipulation.

Findings

Outperforms baselines in success rates and pose accuracy.

Achieves 5.4 FPS inference speed.

Generalizes to unseen objects and transformations.

Abstract

We present RiEMann, an end-to-end near Real-time SE(3)-Equivariant Robot Manipulation imitation learning framework from scene point cloud input. Compared to previous methods that rely on descriptor field matching, RiEMann directly predicts the target poses of objects for manipulation without any object segmentation. RiEMann learns a manipulation task from scratch with 5 to 10 demonstrations, generalizes to unseen SE(3) transformations and instances of target objects, resists visual interference of distracting objects, and follows the near real-time pose change of the target object. The scalable action space of RiEMann facilitates the addition of custom equivariant actions such as the direction of turning the faucet, which makes articulated object manipulation possible for RiEMann. In simulation and real-world 6-DOF robot manipulation experiments, we test RiEMann on 5 categories of manipulation tasks with a total of 25 variants and show that RiEMann outperforms baselines in both task success rates and SE(3) geodesic distance errors on predicted poses (reduced by 68.6%), and achieves a 5.4 frames per second (FPS) network inference speed. Code and video results are available at https://riemann-web.github.io/.