Projects

Autonomous Donkey Car Project

I collaborated on an autonomous car project through the open source DonkeyCar library using a Raspberry Pi, camera, computer vision, and embedded control software to perform lane following, stop-sign detection, and ArUco-based navigation. I created this project for Short Circuit to teach students practical robotics, perception, control, sensor integration, data collection, and autonomous decision-making. I am currently developing the vehicle’s navigation logic and integrating multiple behaviors into a complete self-driving system.

Indoor Path Localization Using Channel State Information
(EECS 498)

Collaborated on the design of a CSI-based indoor localization framework to identify occupant locations within a building during emergencies

Collected data across five distinct locations
Used a neural network model to predict device location from wireless packets
Achieved ~80% accuracy on training/testing data
Generalization to new data proved challenging only ~40% accuracy

This project was valuable as I learned about OFDM modulation and neural network design

Check out our presentation here

DC Programmable Load Project

I collaborated on a DC programmable load project using an Arduino, power MOSFET, current-sensing circuitry, DAC, and custom PI control loop to regulate current in constant-current mode. I created this project for Short Circuit to teach students practical power electronics, PCB design, embedded control, circuit protection, calibration, and closed-loop system tuning. I also developed a custom PCB and expanded the system with thermal, reverse-polarity, and overcurrent protection.

Ball and Beam Project

I built a ball and beam control system using an Arduino Uno with a custom PID controller, integrating a NEMA 17 stepper motor, DRV8825 driver, and HC SR04 ultrasonic sensor. I created this project for my startup to teach students practical control systems design spanning modeling electronics embedded software and controller tuning. I am currently developing a physics based simulation to complement the hardware and support model based testing.

Smartwatch Project

Built a custom smartwatch by developing fully customizable firmware on top of a pre-designed hardware platform. This system integrates an accelerometer, touchscreen display, Wi-Fi, RTOS and power management.

I created this project for my startup to teach students how to design real embedded systems. This covers everything from low-level firmware to multitasking, sensor integration, and UI development. It serves as a hands-on, industry relevant introduction to embedded engineering

Autonomous Robbery Project
(EECS 452)

Designed an autonomous "cop and robber" RC car simulation

Demonstrated image and audio processing for autonomous vehicles
Used binary thresholding and a proportional controller for cop car
Created an amplitude-based audio detection algorithm for robber car

Adaptive Cruise Controller (EECS 461)

Collaborated to develop an embedded adaptive cruise controller with haptic feedback

Designed completely through Simulink
Auto-generated C Code
Run on an NXP microcontroller

This project taught me about real-industry practices with rapid prototyping

Magnetic Levitation Lab (EECS 460)

Collaborated to design a robust controller and deployed it on a magnetic levitation device

The two opposing forces were a gravitational force from the ball's weight and a magnetic force created by an induced current
Linearized system dynamics, calculated equilibrium point, and designed a PID controller with zero steady-state error and a settling time of 3 seconds
Added a pre-compensator to reduce overshoot, computed gain and phase margins, and simulated controller under various nonlinear simulations
Tested our controller and precompensator in a lab where we evaluated our system through multiple simulations
Created report to highlight design, results, and reflections throughout the experiment

You can view our final project report here

Musical Artist Identification (EECS 351)

Collaborated on a MATLAB program designed to identify a music artist based on their own songs, addressing the scenario of hearing a song without knowing the artist

Selected 10 songs from each of four different artists
Tested whether our program could match each track to its performer
Achieved ~77% frame per frame accuracy

This project gave me valuable experience with audio filtering techniques, audio fingerprinting, and machine learning

Check out our website here

Custom RISC ISA Toolchain: Assembler, Simulator, Linker, Runtime (EECS 370)

Built an assembler, simulator, linker, and recursive runtime for a custom RISC ISA, implementing instruction encoding, control flow, memory modeling, and stack-based execution
Created mutation-tested assembly suites that exposed buggy implementations and
Verified correctness, error handling, and architectural edge cases across the toolchain

You can view my projects here

Algorithmic Systems Design and Optimization Projects (EECS 281)

Designed and implemented end-to-end algorithmic systems using advanced data structures (heaps, hash maps, graphs, balanced trees) to solve large-scale search, optimization, and scheduling problems with strict performance constraints
Analyzed runtime tradeoffs and engineered optimized solutions through profiling, asymptotic reasoning, pruning, and careful memory management
Achieved correctness on complex edge cases while meeting time and space limits