Rapid computational design and simulation of highly dynamic legged robots

Sim, Youngwoo

Permalink

https://hdl.handle.net/2142/129458 Copy

Description

Title

Rapid computational design and simulation of highly dynamic legged robots

Author(s)

Sim, Youngwoo

Issue Date

2025-04-30

Director of Research (if dissertation) or Advisor (if thesis)

Ramos, Joao

Doctoral Committee Chair(s)

• Ramos, Joao
• Yim, Justin

Committee Member(s)

• Salapaka, Srinivasa
• Kim, Joohyung

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

Humanoid Design, Actuation Design, Design Optimization

Abstract

The goal of the proposed research is to develop an interactive computational design tool that assists human designers in navigating the high-dimensional design space of robotic actuation systems. Existing design methods and methods face two major challenges: 1) they often suffer from heavy computational demands and slow convergence, and 2) they require a steep learning curve for simulating desired behaviors and validating candidate hardware models. This research aims to accelerate both areas, i.e., reducing design time from 100 hours to 1 hour, to enable interactive design, allowing hardware designers to quickly prototype, visualize and explore new designs. This article proposes a framework comprising three components: behavior synthesis, control-aware design optimization, and simulation-based validation. Three primary academic contributions are anticipated from the development of the proposed framework: it enables 1) a first-principle-based approach to system-level hardware design which used to rely heavily on heuristics, moving it from an art form toward a more rigorous science; 2) greater interactive engagement for mechanical designers in both design and simulation, extending their scope from individual component choices to high-dimensional behaviors; and 3) hardware engineers to gain and extract system-level design intuitions, trade-offs and guidelines.

Graduation Semester

2025-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/129458

Copyright and License Information

Copyright 2025, Youngwoo Sim

Owning Collections