Full-stack autonomous quadrotor system combining computer vision, SLAM, optimal control, reinforcement learning, and waypoint navigation.
Highlights
Implemented computer vision pipelines for object tracking and ArUco-marker-based localization, with optional YOLO-style detectors for semantic awareness.
Used SLAM and mapping to build an environment representation for collision-free waypoint planning.
Combined low-level PID/MPC flight control with a high-level reinforcement-learning policy and waypoint planner for fully autonomous missions.