Nolan Wagener

I received my PhD in Robotics at the Georgia Institute of Technology, where I was advised by Byron Boots (UW) and Panagiotis Tsiotras. I am interested in machine learning for robotics, particularly to increase the capabilities of robots. I am fortunate to have been funded by the NSF Graduate Research Fellowship and received and been nominated for several "best paper" awards at robotics conferences. I have also interned at Microsoft Research in the reinforcement learning group.

I am currently in the industry job market for a position in cutting-edge research in robotics and machine learning.

Email  /  CV  /  GitHub  /  Google Scholar  /  LinkedIn

Thesis  /  Thesis Talk

profile photo

Research

project image

H-GAP: Humanoid Control with a Generalist Planner


Zhengyao Jiang*, Yingchen Xu*, Nolan Wagener, Yicheng Luo, Michael Janner, Edward Grefenstette, Tim Rocktäschel, Yuandong Tian
International Conference on Learning Representations (ICLR), 2024
Spotlight Presentation
Paper  /  Website  /  Code

A single humanoid policy trained on MoCapAct that can be used for many downstream tasks.

project image

TerrainNet: Visual Modeling of Complex Terrain for High-Speed, Off-Road Navigation


Xiangyun Meng, Nathan Hatch, Alexander Lambert, Anqi Li, Nolan Wagener, Matthew Schmittle, JoonHo Lee, Wentao Yuan, Zoey Chen, Samuel Deng, Greg Okopal, Dieter Fox, Byron Boots, Amirreza Shaban
Robotics: Science and Systems (RSS), 2023
Paper  /  Website  /  Talk  /  Poster  /  Slides

A vision-based system for off-road driving that can predict semantic and geometric terrain information from stereo camera input.

project image

MoCapAct: A Multi-Task Dataset for Simulated Humanoid Control


Nolan Wagener, Andrey Kolobov, Felipe Vieira Frujeri, Ricky Loynd, Ching-An Cheng, Matthew Hausknecht
Neural Information Processing Systems (NeurIPS), 2022
Paper  /  Website  /  Blog  /  Talk  /  Short Talk  /  Code  /  Dataset  /  Poster  /  Slides

We release a dataset of high-quality experts and their rollouts for tracking 3.5 hours of MoCap data in dm_control. We use this dataset to train policies that can track the entire dataset, efficiently transfer to other tasks, and perform physics-based motion completion.

project image

DARPA Robotic Autonomy in Complex Environments with Resiliency (RACER)


University of Washington, 2021
Website  /  Article  /  Video

High-speed off-road autonomy in complex terrain. I led development of a control stack that enables a Polaris RZR vehicle to traverse a wide variety of unstructured environments and terrains.

project image

Safe Reinforcement Learning Using Advantage-Based Intervention


Nolan Wagener, Byron Boots, Ching-An Cheng
International Conference on Machine Learning (ICML), 2021
Paper  /  Talk  /  Code  /  Poster  /  Slides

An intervention-based technique for safe reinforcement learning which is based on an advantage function estimate with respect to a given baseline policy. Our work comes with strong theoretical guarantees on performance after training and safety both during and after training, which we corroborate with simulated experiments.

project image

Scalable, Adaptive, and Resilient Autonomy (SARA)


University of Washington, 2020
Video

Perception, planning, and control for navigation though complex, off-road terrain. I developed a control stack for enabling a Clearpath Warthog to navigate through a variety of complex, natural terrain, including mud, vegetation, and snow.

project image

Exploiting Singular Configurations for Controllable, Low-Power Friction Enhancement on Unmanned Ground Vehicles


Adam Foris, Nolan Wagener, Byron Boots, Anirban Mazumdar
IEEE Robotics and Automation Letters (RA-L), 2020
Paper  /  Video  /  Talk

A low-power wheel attachment that greatly increases traction on low-friction surfaces like synthetic ice.

project image

An Online Learning Approach to Model Predictive Control


Nolan Wagener*, Ching-An Cheng*, Jacob Sacks, Byron Boots
Robotics: Science and Systems (RSS), 2019
Winner of Best Student Paper Award
Finalist for Best Systems Paper Award

Paper  /  Video  /  Talk  /  Poster  /  Slides

A connection between model predictive control (MPC) and online learning, demonstrating that many well-known MPC algorithms are special cases of dynamic mirror descent.

project image

Fast Policy Learning Through Imitation and Reinforcement


Ching-An Cheng, Xinyan Yan, Nolan Wagener, Byron Boots
Uncertainty in Artificial Intelligence (UAI), 2018
Plenary Presentation
Paper  /  Talk  /  Poster

A simple yet effective policy optimization algorithm that first performs imitation learning and then switches to reinforcement learning. Theory and experiments show that the policy learns nearly as fast as if performing reinforcement learning starting from the expert that we’re imitating.

project image

Information Theoretic MPC for Model-Based Reinforcement Learning


Grady Williams, Nolan Wagener, Brian Goldfain, Paul Drews, James Rehg, Byron Boots, Evangelos Theodorou
IEEE International Conference on Robotics and Automation (ICRA), 2017
Finalist for Best Conference Paper Award
Paper  /  Video  /  Code  /  Colab  /  Poster  /  Slides

MPPI, a sampling-based model predictive control algorithm, which can handle general nonlinear dynamics and discontinuous costs. We demonstrate MPPI’s capability on an aggressive driving task.

project image

Learning Contact-Rich Manipulation Skills with Guided Policy Search


Sergey Levine, Nolan Wagener, Pieter Abbeel
IEEE International Conference on Robotics and Automation (ICRA), 2015
Winner of Best Robotic Manipulation Paper Award
Paper  /  Article  /  Video  /  Talk

An application of guided policy search, a reinforcement learning algorithm which alternates between trajectory optimization and supervised learning, that allows a PR2 robot to efficiently learn manipulation skills.




Other Projects

These include unpublished research and coursework.

project image

Consistent Dropout for Policy Gradient Reinforcement Learning


Matthew Hausknecht, Nolan Wagener
arXiv, 2022
Paper

Identify a stability issue with naive application of dropout in policy gradient and propose a simple fix for it.

project image

Value Function Learning for AutoRally Racing


CS 8751: Robotics Research, Georgia Tech, 2019
Report  /  Poster

Establish a connection between maximum entropy reinforcement learning and model predictive control, specifically MPPI. Propose a value learning scheme based on MPPI, and demonstrate results on simulated AutoRally task.

project image

End-to-End Training of an Optimal Control Network


CS 7643: Deep Learning, Georgia Tech, 2017
Poster

Perform imitation learning with a policy based on MPPI.

project image

Empirical Evaluation of Recurrent Neural Networks for System Identification


CS 8803 STR: Statistical Techniques in Robotics, Georgia Tech, 2016
Report  /  Slides

Investigate training of recurrent neural networks for system identification, including pre-training techniques for identifying the (unobserved) state.

project image

The Printzerria


ME 102B: Mechatronics Design, UC Berkeley, 2014
Website  /  Video

A 3D printer for making pizza, with “extrusion” tubes for dough, tomato paste, and cheese. I programmed the robot.

project image

Line-Maze Solving Robot


EE 149: Embedded Systems, UC Berkeley, 2013
Report  /  Video  /  Slides

An m3pi robot that can navigate an unknown line maze, complete with doors and waypoints represented by colored tape. I programmed the maze navigation algorithm and the color detection from the color sensor.


Design and source code from Leonid Keselman's website, which itself is based on Jon Barron's website.