This solution utilizes 2D Vision, Deep Learning (Instance Segmentation), Reinforcement Learning (RL), and Motion Planning (OMPL) algorithms to rapidly recognize and sort recyclable plastics on a moving conveyor belt.

The system features high-performance vision capable of quickly identifying objects even with highly irregular shapes. Through advanced motion planning, it executes sorting tasks with exceptional speed and precision.

Effortless Robot Teaching

• Teaching Automation: Shortens setup time by automating vision calibration and robot teaching through deep learning.

• Operational Flexibility: Eliminates the need for manual recalibration or re-teaching even when working conditions change.

• Superior Recognition: High-performance detection of various plastic shapes using deep learning-based object recognition.

• Self-Learning Optimization: Automatically learns the optimal picking position tailored to the specific shape of each product.

Rapid Return on Investment (ROI)

• High-Speed Performance: Capable of ultra-fast operations at a rate of approximately 1 pick per second.

• Operational Optimization: Reduces maintenance costs and optimizes paths through continuous learning, further shortening cycle times.

• Automatic Compensation: Maintains accuracy by automatically correcting movements even if the layout shifts or the robot's mechanical precision deviates due to aging.

• Minimized Downtime: Ensures stable productivity by reducing system failures and manual interventions.