Vision and RTLS Safety Implementation in an Experimental Human-Robot Collaboration Scenario

Area of Science:

• Robotics
• Industrial Automation
• Computer Vision

Background:

• Human-robot collaboration is increasing in industrial settings due to its flexibility.
• Traditional safety measures are often insufficient for advanced collaborative environments.
• Emerging technologies offer new solutions for safe and adaptive robotic workplaces.

Purpose of the Study:

• To propose a novel safe robotic workplace system that adapts its operation and speed based on surrounding environmental stimuli.
• To integrate advanced technologies for enhanced safety and seamless collaboration in industrial settings.
• To reduce operational downtimes in collaborative robotic systems through innovative exception handling.

Main Methods:

• Utilized a depth camera with passive stereo vision to establish dynamic safety zones and detect objects, calculating their distance via pixel shift.
• Implemented the Histogram of Oriented Gradients (HOG) for precise human identification within the workspace.
• Integrated a real-time location system (RTLS) with tags for unequivocal identification of autonomous trolleys supplying materials.

Main Results:

• The system successfully identified objects and humans within safety zones, adjusting the robotic workplace's speed or halting operations accordingly.
• Autonomous trolleys entering safety zones did not disrupt the workplace speed, demonstrating effective exception handling.
• Simulations confirmed the system's compliance with safety measures in various operational scenarios.

Conclusions:

• The integration of an RTLS with a vision-based safety system provides an innovative approach to exception handling in collaborative robotics.
• This interconnected system enhances safety and collaboration while significantly reducing system downtimes.
• The proposed adaptive robotic workplace represents a significant advancement in industrial automation safety and efficiency.