Using Full Pose Measurement for Serial Robot Calibration.

Area of Science:

• Robotics
• Kinematics
• Sensor Fusion

Background:

• Robot operation relies on accurate kinematic models and base-to-world transformations.
• External sensors providing 3D locations or 6D poses are crucial for calibration and registration.
• Current methods often utilize 3D data, potentially limiting accuracy.

Purpose of the Study:

• To compare the effectiveness of 3D location versus full 6D pose measurements for robot calibration.
• To evaluate the impact of measurement type on robot orientation and position errors.
• To determine the optimal sensor data for precise robot kinematic parameter and registration transformation determination.

Main Methods:

• Extensive simulations were conducted on a 7 degrees of freedom robot arm.
• Positional and rotational components of sensor measurements were perturbed with pseudo-noise.
• Calibration and registration procedures using 3D data were compared against those using 6D pose data.

Main Results:

• Full 6D pose measurements significantly reduce robot orientation errors compared to 3D data alone.
• Robot position errors were found to be comparable regardless of whether 3D or 6D pose data was used.
• Simulations demonstrated the superiority of 6D pose data for improving rotational accuracy in robot calibration.

Conclusions:

• Full 6D pose measurements are superior to 3D location data for minimizing robot orientation errors during calibration and registration.
• For precise robot kinematic parameter determination and coordinate frame registration, 6D pose sensing is recommended.
• The findings provide valuable insights for optimizing sensor selection in robotic calibration tasks.