Robotics Engineering student at the University of Michigan focused on embedded robotics software, autonomous systems, and controls.
My experience spans physical mobile-robot autonomy, autonomous-vehicle sensor validation, open-source robotic prostheses, and embedded sensing systems. Recent work includes implementing and validating an end-to-end ROS 2 localization, mapping, and SLAM stack on a physical differential-drive robot.
Implemented a physical ROS 2 autonomy stack spanning embedded wheel control, odometry, waypoint navigation, LiDAR mapping, Monte Carlo Localization, and SLAM.
- Tuned feedforward/PID wheel-speed control to 0.0028 m/s MAE and implemented a C++ rotate-translate-rotate waypoint controller
- Implemented motion-compensated LiDAR processing, Bresenham/log-odds occupancy mapping, probabilistic action and sensor models, and low-variance particle filtering
- Achieved 0.026 m localization RMSE, compared with 0.149 m for odometry alone
- Profiled MCL at approximately 64 ms/update with 400 particles on a Raspberry Pi and tuned real-robot uncertainty to eliminate duplicated map geometry
Contributed to software and development infrastructure supporting the University of Michigan's Open-Source Leg, an open-source robotic prosthesis platform used to develop and test control algorithms on physical hardware.
- Implemented Python-based software improvements for research workflows
- Supported reproducible Raspberry Pi system configuration and developer onboarding
- Contributed merged network-debugging and SMTP setup documentation
- Collaborated through GitHub issues, pull requests, code review, and issue resolution
- Validated BOM and CAD data during a SolidWorks-to-OnShape migration
Contributed to an embedded soft-sensing system that converts a fabric touch matrix into a responsive controller with real-time visual feedback.
- Developed Python and MicroPython components for Raspberry Pi and microcontroller hardware
- Worked with matrix scanning, serial communication, and OLED visualization
- Supported low-latency state updates for a scalable 5x5 touch interface
Earlier independent ROS 2 simulation project exploring autonomous navigation in a custom multi-room office environment.
- Integrated TurtleBot3, Gazebo Classic, RViz2, SLAM Toolbox, AMCL, and Nav2
- Generated 2D LiDAR occupancy maps and demonstrated point-to-point autonomous navigation
| Area | Tools |
|---|---|
| Languages | C, C++, Python, MATLAB |
| Robotics | ROS 2, SLAM, MCL, Nav2, AMCL, LiDAR, Gazebo, RViz2 |
| Embedded and systems | Raspberry Pi, RP2040, Linux, CAN, serial communication, Docker, Git |
| Engineering | Controls, state estimation, sensor integration, hardware-in-the-loop validation |
- Embedded robotics software and controls
- Physical robot localization, mapping, planning, and navigation
- Performance-conscious C++ systems integration