Learning Deep Parameterized Skills from Demonstration for Re-targetable Visuomotor Control

TL;DR

This paper presents a goal-parameterized neural network policy enabling robots to learn visuomotor skills that generalize to unseen targets, demonstrated through simulation and physical robot experiments with significant goal coverage.

Contribution

Introduces an end-to-end goal-parameterized neural network policy that generalizes to unseen goals using limited training data, applicable to various visuomotor tasks.

Findings

Model trained on 33% of goals generalizes to over 90% of targets.

Successfully maps pixel coordinates to robot policies for specific goals.

Effective in both simulation and real robot experiments.

Abstract

Robots need to learn skills that can not only generalize across similar problems but also be directed to a specific goal. Previous methods either train a new skill for every different goal or do not infer the specific target in the presence of multiple goals from visual data. We introduce an end-to-end method that represents targetable visuomotor skills as a goal-parameterized neural network policy. By training on an informative subset of available goals with the associated target parameters, we are able to learn a policy that can zero-shot generalize to previously unseen goals. We evaluate our method in a representative 2D simulation of a button-grid and on both button-pressing and peg-insertion tasks on two different physical arms. We demonstrate that our model trained on 33% of the possible goals is able to generalize to more than 90% of the targets in the scene for both simulation and robot experiments. We also successfully learn a mapping from target pixel coordinates to a robot policy to complete a specified goal.