Robustness via Retrying: Closed-Loop Robotic Manipulation with Self-Supervised Learning

TL;DR

This paper introduces a self-supervised, retry-based approach for robotic manipulation from raw images, enabling robots to learn complex tasks through continuous retrying and object tracking, with minimal labeled data.

Contribution

It proposes a novel self-supervised algorithm for image registration that allows robots to track objects during retries, improving task success despite imperfect models.

Findings

Robots can learn complex manipulation tasks from raw images using only autonomous data.

A self-supervised image registration method enables continuous task retries.

Models trained on 160 hours of data successfully manipulate unseen objects.

Abstract

Prediction is an appealing objective for self-supervised learning of behavioral skills, particularly for autonomous robots. However, effectively utilizing predictive models for control, especially with raw image inputs, poses a number of major challenges. How should the predictions be used? What happens when they are inaccurate? In this paper, we tackle these questions by proposing a method for learning robotic skills from raw image observations, using only autonomously collected experience. We show that even an imperfect model can complete complex tasks if it can continuously retry, but this requires the model to not lose track of the objective (e.g., the object of interest). To enable a robot to continuously retry a task, we devise a self-supervised algorithm for learning image registration, which can keep track of objects of interest for the duration of the trial. We demonstrate that this idea can be combined with a video-prediction based controller to enable complex behaviors to be learned from scratch using only raw visual inputs, including grasping, repositioning objects, and non-prehensile manipulation. Our real-world experiments demonstrate that a model trained with 160 robot hours of autonomously collected, unlabeled data is able to successfully perform complex manipulation tasks with a wide range of objects not seen during training.