About
Christopher Allred
PhD Candidate, AI Robotics Researcher, Software Engineer, Mechanical EIT
Welcome to my website! If your curiosity has brought you here, you’re in the right place. I love all things robotics and I’m passionate about developing innovative solutions to complex problems.
I am a robotics researcher focused on improving terrain cost estimation and dynamic motion learning for real-world robotic systems. My work emphasizes data-centric approaches to boost the performance and adaptability of legged robots. Explore my projects and research, and hopefully, you’ll learn something new about the field of robotics!
Education
| PhD. Computer Science | B.S. Mechanical Engineering |
|---|---|
| 2021 - 2025 | 2016 - 2021 |
| Utah State University | Utah State University |
Experience
Research Fellow
Army Research Lab (ARL): Computational & Information Sciences Directorate
June 2021 - Present
- Lead NIVIDA-ARL collaboration, orchestrating objectives and engineering efforts.
- Developed jumping gaits using reinforcement learning algorithms such as PPO for the Go1 robot in Omniverse Isaac Sim.
- Developed ML algorithm for ARL’s LLAMA quadrupedal platform.
- LSTM classification (96% accuracy) and regression (25.23W RSMSE) predictions on time series terrain data.
- Applied transfer learning to a ResNet50 model for terrain power estimation.
Projects & Technologies:
- Developed Jumping Reinforcement learning algorithm for Go1 quadruped
- Created terrain categorization models on JPL’s LLAMA
- Developed power model for Boston Dynamics legged robot Spot
Publications:
- Detecting Ballistic Motions in Quadruped Robots: A Boosted Tree Motif Classifier for Understanding Reinforcement Learning
- Terrain Dependent Power Estimation for Legged Robots in Unstructured Environment
- Improving Methods for Multi-Terrain Classification Beyond Visual Perception
Research Assistant
Direct Laboratory, Utah State University
Aug 2022 - Aug 2023
- Mentored and supervised undergraduate and master’s students.
- Developed and tested new algorithms for multi-agent robotics teaming research.
- Detected complex motion patterns in reinforcement learning training in Isaac Gym.
Publications:
- Unknown Building Exploration Simulator (UBES)
- Divide & Survey: Observability Through Multi-Drone City Roadway Coverage
Graduate Teaching Assistant
Utah State University
Aug 2021 - May 2023
- Intelligent Systems (2023), Multi-Agent Systems (2022), and Modern C++ (2021).
Software Engineer
BRENKMAN & Company
Mar 2018 - May 2021
- R&D manufacturing and process automation systems.
- Architected and built vision control loop systems in C++ and Python.
- Implemented OpenCV and TensorFlow neural networks for image recognition.
- Automated metal bending and fabrication systems.
Skills and Hobbies
| Category | Details |
|---|---|
| Expertise | Reinforcement Learning, Legged Robotics, Multi-Agent Teaming |
| Languages | Python, C++, Linux |
| Data Science | PyTorch, TensorFlow, CUDA, Scikit-learn |
| Simulation | Isaac Sim Omniverse, ParaView, Unity & Unreal, MuJoCo & Gymnasium, SolidWorks (CAD) |
| Software | Docker, OpenCV |
In my free time, I enjoy a mix of hands-on and outdoor activities. I love 3D printing, playing the piano, and spending time camping when I can. I also like experimenting with tech in my home lab, especially running large language models locally.
One of my exciting moments was winning 1st place at Hackathon 2022, where my team and I developed a project that involved teleoperating a robot arm using virtual reality, which was a challenging but rewarding experience.
Hackathon 2022: 1st place project, Teleoperation of robot arm with VR