Skip to content
Back to Projects
ML / AI 2021-03

Automatic Nerf Remote Sentry

Senior capstone — fully autonomous Nerf turret with OpenCV-based real-time target tracking, motion prediction, and gyroscopic aiming on a Raspberry Pi 4. Trine University Robotics Senior Design (Mar 2021).

Trine Senior Design Capstone
OpenCVComputer VisionRaspberry PiRoboticsPythonMultithreading

What this project proves

Embedded vision systems developer

Autonomous NERF turret system with real-time tracking, prediction, face and gesture workflows, and remote dashboard control.

Core challenge

Keep sensing, prediction, and physical actuation responsive on constrained hardware.

Evaluation lens

Embedded concurrency, prediction logic, and low-latency control.

A complete autonomous system that demonstrates practical CV and control under hardware constraints.

Overview

Designed and built a fully autonomous Nerf turret as part of Trine University’s Robotics Senior Design capstone. The system uses OpenCV for real-time target tracking on a Raspberry Pi 4 with gyroscopic aiming, motion prediction, and a multithreaded control architecture that keeps perception and actuation responsive on constrained hardware.

This project matters because it is a complete embedded vision system, not just a perception demo. Sensing, prediction, aiming, power delivery, and operator override all had to work together under real latency constraints.

What I Owned

  • Built the target-tracking and motion-prediction logic
  • Designed the multithreaded control architecture
  • Integrated the gyroscopic aiming system and servo-control path
  • Helped deliver the hardware + software system as a functional capstone artifact

Hard Problems Solved

  • Low-latency CV on weak hardware: perception had to stay responsive on a Raspberry Pi instead of lab-grade hardware
  • Prediction instead of reaction: simply chasing the previous frame is not enough for moving targets, so the system needed anticipatory aiming
  • Concurrency across sensing and control: vision, prediction, remote I/O, and servo motion had to coexist without blocking each other
  • Hardware/software reliability: custom power and control integration had to be stable enough to support the autonomous behavior

Key Features

  • Real-Time Target Tracking: OpenCV-driven detection with calibration of target shape, size, and color performed live on the Pi.
  • Motion Prediction: Predicts target trajectories for dynamic aiming rather than chasing the last frame.
  • Gyroscopic Aiming System: Custom gyroscope-stabilized servo gimbal for precise yaw/pitch control.
  • Multithreaded Architecture: Splits video processing, prediction, remote control, and servo motion across threads to keep latency low on a low-end device.
  • Remote / Cloud Control: Operator can override or monitor remotely; the same threading model handles network I/O without blocking the CV loop.
  • Custom Power Supply: In-house power-delivery design to drive the servos and Pi reliably from a single battery.

Why It Matters

This project shows practical computer vision under physical and compute constraints: not just detecting objects, but building a full autonomous loop that reacts in real time.

Tech Stack

  • CV: OpenCV, Matplotlib, NumPy
  • Hardware: Raspberry Pi 4 (4 GB), servos, gyroscope, custom PSU
  • Language: Python (multithreaded)
  • Course: Trine University, Robotics Senior Design (Capstone)