Learning Robotic Manipulation Tasks via Task Progress based Gaussian Reward and Loss Adjusted Exploration

TL;DR

This paper introduces a novel deep reinforcement learning approach for robotic manipulation that uses a task progress-based reward and a loss-adjusted exploration policy, achieving state-of-the-art results in complex multi-step tasks.

Contribution

The paper presents RoManNet, a vision-based model architecture, along with TPG reward and LAE policy, to improve learning efficiency and success in multi-step robotic manipulation tasks.

Findings

Outperforms existing methods in success rate and efficiency

Effective in both simulation and real-world environments

Particularly beneficial for multi-block stacking tasks

Abstract

Multi-step manipulation tasks in unstructured environments are extremely challenging for a robot to learn. Such tasks interlace high-level reasoning that consists of the expected states that can be attained to achieve an overall task and low-level reasoning that decides what actions will yield these states. We propose a model-free deep reinforcement learning method to learn multi-step manipulation tasks. We introduce a Robotic Manipulation Network (RoManNet), which is a vision-based model architecture, to learn the action-value functions and predict manipulation action candidates. We define a Task Progress based Gaussian (TPG) reward function that computes the reward based on actions that lead to successful motion primitives and progress towards the overall task goal. To balance the ratio of exploration/exploitation, we introduce a Loss Adjusted Exploration (LAE) policy that determines actions from the action candidates according to the Boltzmann distribution of loss estimates. We demonstrate the effectiveness of our approach by training RoManNet to learn several challenging multi-step robotic manipulation tasks in both simulation and real-world. Experimental results show that our method outperforms the existing methods and achieves state-of-the-art performance in terms of success rate and action efficiency. The ablation studies show that TPG and LAE are especially beneficial for tasks like multiple block stacking. Code is available at: https://github.com/skumra/romannet