Analysis and Case Studies of Nvidia Isaac Gym for Deep Reinforcement Learning-Based Quadruped Robot Control

Journal of Advanced Technology Research, Vol. 9, No. 2, pp. 8-20, Dec. 2024
10.11111/JATR.2024.9.2.008, Full Text:
Keywords: Quadruped Robots, Deep Reinforcement Learning, NVIDIA Isaac Gym, Parallel Simulation, Sim-to-Real Transfer, Robot control
Abstract

This paper provides a comprehensive analysis and survey of existing research on quadruped robot control using deep reinforcement learning, focusing on the NVIDIA Isaac Gym platform. NVIDIA Isaac Gym supports DRL training through parallel simulation and GPU acceleration, enabling quadruped robots to operate stably and efficiently in complex environments. The paper compares the characteristics and performance of various DRL algorithms and analyzes case studies to identify current technological limitations and future research directions.


Statistics
Show / Hide Statistics

Statistics (Cumulative Counts from September 1st, 2017)
Multiple requests among the same browser session are counted as one view.
If you mouse over a chart, the values of data points will be shown.


Cite this article
[IEEE Style]
Y. Choi, C. Ji, Y. Han, "Analysis and Case Studies of Nvidia Isaac Gym for Deep Reinforcement Learning-Based Quadruped Robot Control," Journal of Advanced Technology Research, vol. 9, no. 2, pp. 8-20, 2024. DOI: 10.11111/JATR.2024.9.2.008.

[ACM Style]
Yohan Choi, Changhun Ji, and Youn-Hee Han. 2024. Analysis and Case Studies of Nvidia Isaac Gym for Deep Reinforcement Learning-Based Quadruped Robot Control. Journal of Advanced Technology Research, 9, 2, (2024), 8-20. DOI: 10.11111/JATR.2024.9.2.008.