Why Calibration Matters
A 2mm calibration error at the wrist compounds to 5–10mm at the fingertips. For precision tasks like insertion, peg-in-hole, or small object grasping, this makes the difference between success and failure. Calibration also ensures that data collected across different sessions is spatially consistent — essential for training manipulation policies.
Hand-Eye Calibration
Hand-eye calibration determines the transform between a camera and the robot's end-effector (eye-in-hand) or base (eye-to-hand). Use an AprilTag or ChArUco board, move the robot to 15–20 diverse poses, solve the AX=XB equation. Tools: easy_handeye (ROS), OpenCV, VISP. Validate by commanding the robot to touch a known point and measuring the error.
Maintaining Calibration Over Time
Calibration drifts due to mechanical wear, temperature changes, and accidental bumps. Implement a quick daily check: command the robot to a reference pose and verify alignment with a physical marker. Recalibrate monthly or after any hardware change. SVRC's OpenArm includes a built-in calibration verification routine.