Tech Stack: Python, OpenCV, YOLOv8, Kalman Filter, PID Control, CAD A 2-axis predictive tracking gimbal system developed for Unmanned Ground Vehicles (UGVs) during PuneSymbiHackathon Startupcon 5.0. This system projects a real-time Cartesian coordinate system onto a camera feed and calculates precise pan/tilt servo angles to keep dynamic targets dead-centered. Key Highlights: • Predictive Tracking: Implemented Kalman filtering to calculate object trajectories, allowing the system to actively predict where a target is if line-of-sight is lost over a bump or obstacle. • Human-AI Shared Control: Engineered a seamless transition between autonomous AI tracking and manual teleoperation, allowing an operator to instantly take over via keyboard and auto-lock the second they let go. • Hardware-Ready: Designed a fully 3D-printable pan-tilt mechanism (CAD renders included) and developed a closed-loop PID controller tuned to prevent derivative explosions from UGV chassis vibrations.