Accuracy evaluation of hand-eye calibration techniques for vision-guided robots.

Area of Science:

• Robotics
• Computer Vision
• Calibration Techniques

Background:

• Accurate hand-eye calibration is vital for vision-guided robotic applications such as assembly and inspection.
• Existing calibration methods often assess noise effects collectively, overlooking individual impacts of rotation and translation.
• The influence of robot motion range during calibration is frequently underestimated.

Purpose of the Study:

• To investigate the isolated effects of rotation and translation noise on hand-eye calibration accuracy.
• To evaluate the impact of robot motion range (rotation and translation) on calibration performance.
• To comparatively analyze the performance of six common hand-eye calibration algorithms under varying noise and motion conditions.

Main Methods:

• Quantitative evaluation of six hand-eye calibration algorithms.
• Simulation case studies to assess performance under controlled noise and motion parameters.
• Experimental validation using Universal Robot's UR5e physical robots.

Main Results:

• Algorithms exhibit varied responses to isolated rotation and translation noise; simultaneous methods resist rotation noise, while separate methods handle translation noise better.
• Increasing robot rotation motion range improves separate methods' accuracy but degrades simultaneous methods'.
• Increasing translation motion range enhances simultaneous methods' accuracy but harms separate methods' accuracy.

Conclusions:

• The performance of hand-eye calibration algorithms is sensitive to the specific type and magnitude of noise and the robot's motion range during calibration.
• Algorithm selection and calibration process design should account for isolated noise effects and motion parameters for optimal accuracy.
• Findings provide critical insights for benchmarking algorithms and performing precise hand-eye calibration in robotic systems.