About Us
The Computational methods for Curious Robots Lab is a robotics research group focused on the development of rigorous mathematical, algorithmic, and optimization methods for robotic learning and control led by Dr. Ian Abraham.
Our research breakthroughs are deployed on a wide range of diverse robotic systems to expand their utility and overall success in the most extreme and remote environments. Check out our ongoing projects below!
Ongoing Research Projects
Whole-Body Control for Humanoid & Legged Robots
We develop whole-body control methods that enable robots to perform agile and dexterous motor skills. Our approach combines optimization and learning to achieve robust and efficient robot behavior across diverse platforms and applications.
- Learning temporal-awareness in control policies
- Sample-based whole-body control
- Collaborative whole-body loco-manipulation
- Information theory for contact planning
Marine/Ocean Robotics
We develop optimization methods for long-term autonomy in marine and ocean environments. Our approach combines GPU accelerated programming and algorithmic development, adaptive-time planning and control, and concepts from variational inference, and optimal transport to achieve robust autonomy for underwater robotic systems.
- Optimization for long-term autonomy
- Robust optimization via inference
- Distributed/GPU accelerated autonomous planning
Algorithm Development & Reinforcement Learning
We develop efficient algorithms for reinforcement learning and robotic control. Our approach integrates novel techniques from statistical analysis, trajectory optimization, and hybrid control to enable efficient policy optimization and training.
- Stochastic smoothing for non-differentiable learning
- Fast and reliable policy optimization
- Decoupled policy gradients for learning vision-to-action policies
Interested in Joining?
Those interested in joining our group are expected to have strong competency is programming, optimization, systems and control theory, kinematics and dynamics of robotic systems, and rigorous mathematical theory and proof writing. Candidates should have a strong curiosity and drive to develop algorithms to expand the capabilities of robots, analyze their performance, and prove reliability through mathematical rigor and repeated experimentation.
Other useful skills:
- ROS 1/2 or LCM experience
- Extensive Python and C/C++ coding experience
- Exposure to modern ML/AI autodiff libraries like JAX/PyTorch
- Proof writing abilities
If this research sounds exciting then please fill out the following form and detail how you would contribute to the CoCuRo Lab.