Relevant skills: Python, Drake, CasADi, Robot Kinematics and Dynamics, Optimal Control, Trajectory Optimization, Docker, Quaternions
Investigate and implement nonlinear trajectory optimization techniques for planning and control of a quadcopter using Drake.
Relevant skills: ROS, MoveIt, Klamp’t, isaacgym, C++, Python, Path Planning, Docker
The Image Guided Medical Robotics (IGMR) Lab at the University of Michigan is interested in developing a device that can autonomously manipulate an optical coherence tomographer (OCT) to image retinas. I work with industry standard frameworks, such as ROS2 and MoveIt, to implement the motion planning for this project.
Relevant skills: lcm, C++, Motor Control, Raspberry Pi, SLAM, Path Planning
Program a differential wheeled robot to autonomously navigate through mazes. This project was completed for ROB550 at the University of Michigan Ann Arbor.
Relevant skills: OpenCV, Python, ROS, Robot Kinematics and Dynamics
Program a 5DOF arm to complete pick and place challenges. This project was completed for ROB550 at the University of Michigan Ann Arbor.
Relevant skills: Model Matching, Arduino, Python, Embedded Control, Kalman Filter, UART
The UCSF BioRobotics lab is interested in determining the efficacy of robot powered exoskeletons and prosthetics in the application of physical therapy. To study the instruments’ effects, patients will be instructed to perform certain tasks while undergoing a magnetoencephalography. This project goes over the design of an exo compatible with such magnetically sensitive equipment.
Relevant skills: OnShape, 3D Printing, MicroPython
Soft robotic crawlers are desirable for their adaptability in unstructured environments. We were tasked to build such a robot to autonomously navigate a pipeline, modeled after the competition in RoboSoft 2023.
Relevant skills: PyTorch, Optimal Control
NeuralODE utilizes PyTorch to learn the dynamics of the Franka Emika Panda arm for a planar pushing task. This project was completed for Robot Learning and Planning at the University of Michigan Ann Arbor.
Relevant skills: Optimal Control, GNC, MATLAB
The optimal guidance law for missile pursuit involves solving an optimization that can be computationally expensive. This implementation makes use of motion primitives to approximate the behavior of the optimal guidance law for a 2D evader-pursuer scenario in MATLAB. This project was completed for AEROSP584:Navigation and Guidance from Perception to Control at the University of Michigan Ann Arbor.