Improved Learning of Robot Manipulation Tasks via Tactile Intrinsic Motivation

TL;DR

This paper introduces a tactile-based intrinsic motivation and contact-prioritized experience replay to improve exploration in deep reinforcement learning for robotic manipulation, leading to faster learning and better performance on benchmarks.

Contribution

It proposes a novel tactile intrinsic reward and contact-focused experience replay, enhancing exploration efficiency in robotic manipulation tasks.

Findings

Accelerated learning in manipulation tasks

Outperforms state-of-the-art methods

Effective in sparse reward settings

Abstract

In this paper we address the challenge of exploration in deep reinforcement learning for robotic manipulation tasks. In sparse goal settings, an agent does not receive any positive feedback until randomly achieving the goal, which becomes infeasible for longer control sequences. Inspired by touch-based exploration observed in children, we formulate an intrinsic reward based on the sum of forces between a robot's force sensors and manipulation objects that encourages physical interaction. Furthermore, we introduce contact-prioritized experience replay, a sampling scheme that prioritizes contact rich episodes and transitions. We show that our solution accelerates the exploration and outperforms state-of-the-art methods on three fundamental robot manipulation benchmarks.