Predicting Flying Robot Dynamics with Deep Learning

Authors

  • Brian Li Henry M Gunn High School
  • Nathan Lambert Mentor, University of California, Berkeley

DOI:

https://doi.org/10.47611/jsrhs.v10i3.1839

Keywords:

Machine Learning, Robotics, Robot Dynamics, Prediction

Abstract

With the rapid increase in the power of computing and technological advances in robotics, research in the field of robotics has rapidly become very expansive. Being able to accurately predict movements of a robot is vital to many applications within this field, allowing for more precise simulation and prototyping as well as more accurate control of robotic systems. In this paper, we present an adaptable neural network that accurately predicts the movement of quadcopter robotic agents which can be expanded to encompass many more robots and applications given the requisite data, producing accurate results within a small margin for error.

Downloads

Download data is not yet available.

Author Biographies

Brian Li, Henry M Gunn High School

Brian Li is a rising Senior at Henry M. Gunn High School.  He has been passionate about Computer Science from a young age, and has been working on many CS related projects throughout High School. He has a strong interest in Machine Learning. He believes there is extreme potential in Artificial Intelligence to make the world a better place and hopes to perform more research in the future. He has had one paper published and one under review. He has also been volunteering with StreetCode Academy to help young kids develop their appreciation and knowledge of CS.

Nathan Lambert, Mentor, University of California, Berkeley

Nathan Lambert is  a PhD Candidate at the University of California, Berkeley working at the intersection of machine learning and robotics. He is a member of the Department of Electrical Engineering and Computer Sciences, advised by Professor Kristofer Pister in the Berkeley Autonomous Microsystems Lab. Nathan has worked extensively with Roberto Calandra at Facebook AI Research and is joining DeepMind Robotics remotely for the summer of 2021. During his Ph.D., he was awarded the UC Berkeley EECS Demetri Angelakos Memorial Achievement Award for Altruism.

References or Bibliography

Spot | Boston Dynamics. https://www.bostondynamics.com/spot. Accessed 8 Apr. 2021.

Atlas® | Boston Dynamics. https://www.bostondynamics.com/atlas. Accessed 8 Apr. 2021.

Smith, J.O. Physical Audio Signal Processing, http://ccrma.stanford.edu/~jos/pasp/, online book, 2010 edition, accessed 7 Apr. 2021.

Nanyang Technological University. "Scientists develop 'mini-brains' to help robots recognize pain and to self-repair." ScienceDaily. ScienceDaily, 15 October 2020.

https://www.sciencedaily.com/releases/2020/10/201015101812.htm.

Drew, Daniel S., et al. “Toward Controlled Flight of the Ionocraft: A Flying Microrobot Using Electrohydrodynamic Thrust With Onboard Sensing and No Moving Parts.” IEEE Robotics and Automation Letters, vol. 3, no. 4, Oct. 2018, pp. 2807–13. DOI.org (Crossref), doi:10.1109/LRA.2018.2844461.

Lambert, Nathan. Natolambert/Dynamicslearn. 2018. 2021. GitHub,

https://github.com/natolambert/dynamicslearn.

Lambert, Nathan O., et al. “Learning Accurate Long-Term Dynamics for Model-Based Reinforcement Learning.” ArXiv:2012.09156 [Cs], Dec. 2020. arXiv.org, http://arxiv.org/abs/2012.09156.

NN SVG. http://alexlenail.me/NN-SVG/index.html. Accessed 8 Apr. 2021.

Published

11-22-2021

How to Cite

Li, B., & Lambert, N. (2021). Predicting Flying Robot Dynamics with Deep Learning. Journal of Student Research, 10(3). https://doi.org/10.47611/jsrhs.v10i3.1839

Issue

Vol. 10 No. 3 (2021)

Section

HS Research Articles

Copyright (c) 2021 Brian Li; Nathan Lambert

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Copyright holder(s) granted JSR a perpetual, non-exclusive license to distriute & display this article.