MAR-XXV Robot Controller: Closed-Loop AI Navigation

Computer Vision | Embedded Systems | Robotics

This project delivers a responsive, AI-driven control system for the MAR-XXV mobile robot platform. It integrates onboard computer vision for object detection with a low-latency web interface for teleoperation and telemetry streaming.

System Architecture

• Brain: Raspberry Pi 4 running a multi-threaded Python control service.
• Vision: OpenCV-based pipeline for real-time lane tracking and object recognition (Haar Cascades/YOLO).
• Control Interface: FastAPI backend serving a React-based dashboard for live video feed (WebSocket) and joystick control.
• Actuation: Arduino slave controller handling PWM motor drive and encoder feedback via serial communication.

Key Capabilities

• Autonomous Navigation: Lane following and obstacle avoidance logic.
• Live Streaming: Low-latency MJPEG stream over local network.
• Remote Control: Tablet-friendly web interface for manual override.

Repository Structure

• /server: FastAPI backend and video processing services.
• /client: React frontend for control dashboard.
• /firmware: Arduino sketches for motor control.

---

Part of the Computational Textiles portfolio.