Ball Tracker Demo – Distributed Vision-Based Control System

This demo showcases a vision-based ball tracking system using a camera mounted above a tilting platform. The camera captures the position of the ball in real time, and the processed output is displayed on an external monitor for clear visualization.

The ball tracker system is a distributed control application implemented using two Box PCs: the ROScube RQP-T35 (x86) as the vision processing unit and the MXA-200 (ARM) as the control unit.

A vision camera captures real-time image data of a ball on a tilting plate. The ROScube RQP-T35 processes this data to detect and track the ball position at high frame rates. A PID control algorithm computes the required tilt angles to maintain stability or guide the ball.

Control signals are transmitted over standard packet-based Ethernet (non-TSN) to the MXA-200. The control unit generates PWM outputs to drive servo motors, which adjust the plate orientation accordingly.

The system demonstrates:

• Real-time image processing and object tracking
• Closed-loop PID control
• Distributed architecture across heterogeneous platforms (ARM + x86)
• Reliable communication over conventional Ethernet

This makes it a practical reference design for industrial automation, robotics, and motion control systems.

Exhibitor: ADLINK Technology & Alcom Electronics