Projects

RL Target Acquisition and Obstacle Avoidance

Used reinforcement learning to control an agent for locating targets while avoiding dynamic obstacles in a simulated environment. The project poster is available here.

GitHub

RL-Based RC Car Driver

Trained an RC car to follow a track using reinforcement learning with camera input.

GitHub

FPV Drone Simulator

Drone simulator built using OpenSceneGraph and C++ with physical quadcopter modeling and controller input support.

GitHub

Long-range FPV Airplane

Long-range FPV autonomous airplane equipped with a GPS antenna for waypoint missions and return-to-home functionality. DIY foam board airframe with a flight controller running INAV autopilot software. Long-range radio control using ExpressLRS and analog video transmission. Personally constructed the airframe, assembled electronics, tuned controller gains, and troubleshot system issues.

GPU-Accelerated Boids Simulation

Simulated boids flocking behavior on both CPU and GPU using CUDA for parallel speedups.

GitHub

Self-Driving Car with Segmentation

Steered a car through hallways using a segmentation network to identify walls and floor regions.

GitHub

Baseball Catcher

Used stereo vision and kinematics to predict a baseball’s landing location and actuate a net to catch it.

GitHub

Monocular Speed Estimation

Estimated the speed of vehicles using only a single calibrated camera input.

GitHub

UAV Landing Trajectory Generator

Generated landing paths using B-splines and differential flatness-based optimization.

Continuous-Time Batch State Estimation with GPR

Used Gaussian Process Regression to estimate smooth continuous-time trajectories from noisy measurements.

2D ICP for Lidar Scan Matching

Implemented the Iterative Closest Point algorithm for aligning sequential 2D lidar scans and improving pose estimation.