Edinburgh Research Explorer

Emergence of Human-comparable Balancing Behaviors by Deep Reinforcement Learning

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Related Edinburgh Organisations

Access status

Open

Documents

http://ieeexplore.ieee.org/document/8246900/
Original languageEnglish
Title of host publication2017 IEEE-RAS International Conference on Humanoid Robots
PublisherIEEE
Number of pages8
ISBN (Electronic)978-1-5386-4678-6
ISBN (Print)978-1-5386-4679-3
DOIs
StatePublished - 8 Jan 2018
EventIEEE-RAS International Conference on Humanoid Robots - Birmingham, United Kingdom

Conference

ConferenceIEEE-RAS International Conference on Humanoid Robots
CountryUnited Kingdom
CityBirmingham
Period15/11/1717/11/17
Internet address

Abstract

This paper presents a hierarchical framework based on deep reinforcement learning that learns a diversity of policies for humanoid balance control. conventional zero moment point based controllers perform limited actions during under-actuation, whereas the proposed framework can perform human-like balancing behaviors such as active push-off of ankles. The learning is done through the design of an explainable reward based on physical constraints. The simulated results are presented and analyzed. The successful emergence of humanlike behaviors through deep reinforcement learning proves the feasibility of using an AI-based approach for learning humanoid balancing control in a unified framework.

Event

IEEE-RAS International Conference on Humanoid Robots

15/11/1717/11/17

Birmingham, United Kingdom

Event: Conference

Download statistics

No data available

ID: 44599967