Fault-Aware Robust Control via Adversarial Reinforcement Learning

11/17/2020
by   Fan Yang, et al.
11

Robots have limited adaptation ability compared to humans and animals in the case of damage. However, robot damages are prevalent in real-world applications, especially for robots deployed in extreme environments. The fragility of robots greatly limits their widespread application. We propose an adversarial reinforcement learning framework, which significantly increases robot robustness over joint damage cases in both manipulation tasks and locomotion tasks. The agent is trained iteratively under the joint damage cases where it has poor performance. We validate our algorithm on a three-fingered robot hand and a quadruped robot. Our algorithm can be trained only in simulation and directly deployed on a real robot without any fine-tuning. It also demonstrates exceeding success rates over arbitrary joint damage cases.

READ FULL TEXT

page 1

page 3

page 5

page 6

research
10/11/2021

Legged Robots that Keep on Learning: Fine-Tuning Locomotion Policies in the Real World

Legged robots are physically capable of traversing a wide range of chall...
research
02/26/2020

Efficient reinforcement learning control for continuum robots based on Inexplicit Prior Knowledge

Compared to rigid robots that are often studied in reinforcement learnin...
research
10/02/2019

Deep Reinforcement Learning for Single-Shot Diagnosis and Adaptation in Damaged Robots

Robotics has proved to be an indispensable tool in many industrial as we...
research
05/22/2019

Automated shapeshifting for function recovery in damaged robots

A robot's mechanical parts routinely wear out from normal functioning an...
research
11/15/2022

Omnidirectional robot modeling and simulation

A robot simulation system is a basic need for any robotics application. ...
research
10/05/2016

Towards semi-episodic learning for robot damage recovery

The recently introduced Intelligent Trial and Error algorithm (IT&E) ena...
research
07/08/2021

RMA: Rapid Motor Adaptation for Legged Robots

Successful real-world deployment of legged robots would require them to ...

Please sign up or login with your details

Forgot password? Click here to reset