Deep reinforcement learning for quadrupeds

This work presents a thesis on understanding different methods and frameworks developed for Deep Reinforce- ment Learning, and implementing procedure and methods outlined in [1] to develop a control policy for the Stanford Pupper quadruped robot [2]. The project involves simulating the Augmented Random Search (ARS) policy and DeepRL algorithm framework [1] using PyBullet to optimize the movement of the quadruped. The effect of different parameters of the ARS policy and DeepRL algorithm is studied in simulation to evaluate the outcomes and compare their performance and effectiveness.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Systems & Entrepreneurial Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Dashpute, Chinmay

Contributors dc:contributor

• Sreenivas, Ramavarapu S

• Quadrupeds
• Deep Reinforcement Learning

Statement dc:rights

• Copyright 2023 Chinmay Dashpute

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/121555

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F121555