San Francisco Bay Area
At ANL, I design and deploy advanced robotic systems for hazardous and nuclear environments, with a focus on real-time teleoperation, digital twins, and reinforcement learning. Key Contributions: • Designed and deployed bidirectional teleoperation systems using Franka Emika, UR5/UR16e, Kinova Gen3, KUKA LBR iiwa, and custom multi-finger grippers, enabling cross-continental manipulation (U.S. ↔ South Korea) with <20 ms latency via VPN and FastDDS. • Engineered a 4-channel bilateral teleoperation architecture (admittance + impedance control) on a real-time Linux kernel at 800 Hz, with torque/force control, gravity and energy-based friction compensation – reducing joint friction by ~80% and improving operator accuracy, stability, and fatigue resistance. • Developed a Learning from Demonstration (LfD) pipeline combining DMPs and GMM/GMR, achieving ~90% success across varied conditions in hardware validation tasks (e.g., power plug insertion, compliant manipulation). • Built a digital twin and AR/VR framework in Unity and NVIDIA Isaac Sim, modeling contact-rich physics environments and integrating with ROS 2 for hardware-in-the-loop operation, immersive operator training, and high-fidelity visualization. • Integrated HaptX haptic gloves to render real and simulated contact forces, enabling realistic tactile feedback for compliant manipulation in hazardous environments. • Applied NVIDIA Isaac Lab for reinforcement learning with large-scale vectorized physics simulation and domain randomization, accelerating policy training for manipulation tasks. • Designed robust safety and stability mechanisms (workspace/joint/force/torque limits, time-domain passivity control), ensuring reliability of robotic operations in nuclear and safety-critical environments. • Developed and deployed systems using ROS 2, C++, Python, Docker, leveraging real-time kernels and libraries (Eigen, KDL) for advanced dynamics and control.
• Designed and implemented trajectory planning algorithms for KUKA robotic arm, integrating ROS, RViz, Unity Engine, and Microsoft Mixed Reality for real-time control and visualization. • Built a mixed-reality digital twin of the KUKA arm in Unity, deployed on Microsoft HoloLens 2, enabling interactive admittance control and immersive operator interaction. • Developed admittance control pipeline where force/torque sensor feedback was processed in ROS and transmitted to the physical KUKA arm (via Fast Research Interface over UDP) and its mixed-reality counterpart (via Unity Robotics Hub). • Technologies used: ROS/MoveIt, URDF, Microsoft HoloLens 2, Unity Engine, Fast Research Interface (FRI).
• Designed, fabricated, and deployed an Autonomous Underwater Vehicle (AUV) in collaboration with Jacobs University, Germany, sponsored by DAAD. Built a carbon-fiber pressure hull rated to 140 psi, waterproof electronics housing, and customized BlueRobotics T200 thrusters with 3D-printed vanes for enhanced maneuverability. • Developed a fully onboard visual–inertial SLAM system in ROS, integrating Intel RealSense D435i and NVIDIA Jetson for sparse underwater environments, with real-time mapping and video transmission to a ground station via RViz. • Published peer-reviewed research (ECMR 2021, Bonn, Germany) and presented at the 18th IEEE International Workshop on Field & Assistive Robotics (FAR 2021). • Conducted hands-on training workshops on ROS: From Design to Simulation for undergraduate and master’s students, mentoring early-career researchers in robotics development. • Collaborated on interdisciplinary projects, including applying neural networks for plant disease detection using computer vision. Technologies used: ROS, MAVLink, NVIDIA Jetson, Intel RealSense, ANSYS Fluent, SolidWorks, embedded Linux, SLAM, underwater navigation & depth mapping.
• Managed a 40-member student team in the design, simulation, and fabrication of a carrier-operations UAV with wing-folding mechanism for the AIAA Design/Build/Fly international competition. • Directed end-to-end development: aerodynamic modeling, structural fabrication, avionics integration, and embedded systems design. • Developed a MATLAB/Simulink flight dynamics simulator to validate aircraft stability and mission performance before testing. • Coordinated project documentation, technical reviews, and sponsor communications, ensuring on-time competition readiness.
• Led a 15-member design sub-team for aircraft simulation and design tasks across two DBF competition cycles. • Authored technical proposals and final reports submitted to AIAA, demonstrating compliance with mission requirements. • Collaborated with fabrication and electronics teams to transition conceptual designs into competition-ready prototypes.