Learning for a Robot: Deep Reinforcement Learning, Imitation Learning, Transfer Learning

Zhaojie Ju,Liangcai Zeng,Gongfa Li,Jiang Hua

doi:10.3390/s21041278

Zhaojie Ju, Liangcai Zeng + Show 2 more

Open Access

https://doi.org/10.3390/s21041278

Copy DOI

Abstract

Dexterous manipulation of the robot is an important part of realizing intelligence, but manipulators can only perform simple tasks such as sorting and packing in a structured environment. In view of the existing problem, this paper presents a state-of-the-art survey on an intelligent robot with the capability of autonomous deciding and learning. The paper first reviews the main achievements and research of the robot, which were mainly based on the breakthrough of automatic control and hardware in mechanics. With the evolution of artificial intelligence, many pieces of research have made further progresses in adaptive and robust control. The survey reveals that the latest research in deep learning and reinforcement learning has paved the way for highly complex tasks to be performed by robots. Furthermore, deep reinforcement learning, imitation learning, and transfer learning in robot control are discussed in detail. Finally, major achievements based on these methods are summarized and analyzed thoroughly, and future research challenges are proposed.

Highlights

The concept of robot gripping originated in 1962 with industrial robot Unimate which used a two-finger to grab wooden blocks and stack them together
The survey reveals that the latest research in deep learning and reinforcement learning has paved the way for highly complex tasks to be performed by robots
Baram et al proposed the algorithm of model-based generative adversarial imitation learning (MGAI) based on a forward model to make the calculations completely divisible, which allows the use of accurate discriminator gradients to train strategies [94]

Summary

Introduction

The concept of robot gripping originated in 1962 with industrial robot Unimate which used a two-finger to grab wooden blocks and stack them together. The robot is designed to mimic the function of humans, so the pioneers of the field have done a lot of research on the grasp and manipulation mechanism. Human beings can manipulate objects and explore the world in various environments, so we want robots to be as capable as humans. Manipulation of the robot is not as simple as we think even though studies have been conducted for decades [1]. Robotics has gained vast progress in mechanical design, perception, and robust control targeted to grasp and handle objects, robotic manipulation is still a poor proxy for human dexterity. No robots can hand-wash dishes, button a shirt, or peel a potato

Methods

Discussion

Conclusion