I am deeply engaged in several impactful projects that align with advancing robotics and machine learning:
-
Path Planning Contributions to Open Source: Contributing to the open-source repository AutonomousVehicleControlBeginnersGuide, where I am implementing path planning algorithms like A* to make foundational concepts in self-driving car technology more accessible to students and enthusiasts.
-
Reinforcement Learning with MuJoCo: Developing reinforcement learning frameworks using MuJoCo for robotic control. This includes creating robust environments and algorithms to refine robotic motion through advanced reinforcement learning techniques.
-
Imitation Learning for Aloha Kits: Leveraging Action Chunking Transformers to train Aloha Kits for imitation learning tasks. My work focuses on designing policies, generating sample trajectories, and optimizing models to teach robots complex tasks effectively.
-
Sim-to-Real Transfer for Aloha Kits: Building an end-to-end framework to seamlessly transition policies trained in simulation to physical robots, ensuring real-world applicability and performance consistency.
-
Integration of Aloha Kits with LeRobot: Collaborating with Hugging Face to integrate Aloha Kits into the LeRobot framework. This includes enabling multi-camera support and creating streamlined workflows for dataset collection and training.