Date of Award

2023

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical and Aerospace Engineering

Committee Chair

Farbod Fahimi

Committee Member

Yuri Shtessel

Committee Member

Emil Jovanov

Committee Member

Gang Wang

Committee Member

Chang-kwon Kang

Subject(s)

Androids--Control, Bipedalism, Gait in humans--Simulation methods, Reinforcement learning

Abstract

This dissertation examines the use of model-free control methods in various aspects of effective balance and gait control for two-legged humanoid robots. It explores what it should mean to improve balance and gait control for robots and proposes gait fragility as an alternative framework to the prevailing yet undefined dynamic stability objective. Next, it presents techniques for trajectory tracking control without deriving dynamic models, using modifications of the conventional reinforcement learning approach. Finally, it explores the use of deep online reinforcement learning with disturbance and assistance curricula for learning standing push recovery in bipeds.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.