Goal-Conditioned End-to-End Visuomotor Control for Versatile Skill Primitives

Abstract

Visuomotor control (VMC) is an effective means of achieving basic manipulation tasks such as pushing or pick-and-place from raw images. Conditioning VMC on desired goal states is a promising way of achieving versatile skill primitives. However, common conditioning schemes either rely on task-specific fine tuning (e.g. using metalearning) or on sampling approaches using a forward model of scene dynamics i.e. modelpredictive control, leaving deployability and planning horizon severely limited. In this paper we propose a conditioning scheme which avoids these pitfalls by learning the controller and its conditioning in an end-to-end manner. Our model predicts complex action sequences based directly on a dynamic image representation of the robot motion and the distance to a given target observation. In contrast to related works, this enables our approach to efficiently perform complex pushing and pick-and-place tasks from raw image observations without predefined control primitives. We report significant improvements in task success over a representative model-predictive controller and also demonstrate our model's generalisation capabilities in challenging, unseen tasks handling unfamiliar objects.

Efficient Skill Acquisition for Complex Manipulation Tasks in Obstructed Environments

“…However, due to the sample inefficiency of sparse rewards, studies have been primarily conducted in simulated environments or within limited task spaces in the real world. Learning from demonstration (LfD) [20], [21], [22] is an alternative method for a robot to learn manipulation tasks by imitating behaviour in expert demonstrations collected by a human operator, but it often requires a large number of demonstrations to acquire manipulation skills. While Inser-tionNet [23] enables a robot to solve industrial insertion tasks within the entire operational space of a robot manipulator from a small number of demonstrations, it is evaluated in a clean environment without obstruction.…”