Learning Mobile Manipulation through Deep Reinforcement Learning.

Cong Wang,Yvan Petillot,Shuo Li,Sen Wang,Qiyan Tian,Xiaohui Wang,David Lane,Qifeng Zhang

doi:10.3390/s20030939

Abstract

Mobile manipulation has a broad range of applications in robotics. However, it is usually more challenging than fixed-base manipulation due to the complex coordination of a mobile base and a manipulator. Although recent works have demonstrated that deep reinforcement learning is a powerful technique for fixed-base manipulation tasks, most of them are not applicable to mobile manipulation. This paper investigates how to leverage deep reinforcement learning to tackle whole-body mobile manipulation tasks in unstructured environments using only on-board sensors. A novel mobile manipulation system which integrates the state-of-the-art deep reinforcement learning algorithms with visual perception is proposed. It has an efficient framework decoupling visual perception from the deep reinforcement learning control, which enables its generalization from simulation training to real-world testing. Extensive simulation and experiment results show that the proposed mobile manipulation system is able to grasp different types of objects autonomously in various simulation and real-world scenarios, verifying the effectiveness of the proposed mobile manipulation system.

Highlights

Robot manipulation, as one of the most fundamental and challenging research topics in robotics, has attracted significant interest in last decades
We propose a novel mobile manipulation framework based on deep reinforcement learning
The training results of the proposed mobile manipulation system is given in Figure 6, including the maximum and mean rewards and success rates

Summary

Introduction

As one of the most fundamental and challenging research topics in robotics, has attracted significant interest in last decades. Based on the traditional dynamic control techniques, industrial robot manipulators can perform tasks repeatedly with high precision. Most of the existing manipulation systems are fixed in structured environments with no or limited perception capability. They are not adequate for a broad range of tasks and applications in practice, which requires reliable operation in unstructured and dynamic environments. The deep reinforcement learning method has become a new enabler for complex tasks, which are challenging to accomplish for the traditional methods. In the framework of reinforcement learning, the policy π ( a|s) predicts an action a ∈ A based on a state s ∈ S and a reward r ∈ R is received after action.

Methods

Results

Conclusion