Absolute Positioning Accuracy

How far a robot's actual end-effector position ends up from the theoretical target position it was commanded to reach.

Absolute positioning accuracy measures how close the end effector's actual position comes to the theoretical position implied by the command; the ISO 9283 standard calls this pose accuracy (AP). It's distinct from repeatability (RP) — how tightly the robot returns to the same point over repeated trials. Industrial arms are typically 'very repeatable but not very accurate,' because manufacturing tolerances in link lengths and joint zero offsets, together with deflection under load and temperature effects, cause the controller's internal kinematic model to drift away from the real machine. Hand-guiding taught points with a teach pendant only requires good repeatability, but offline programming — or sending coordinates computed in simulation or by a camera straight to the robot — is instead bottlenecked by absolute accuracy. The usual fix is kinematic calibration: measuring the true pose with equipment such as a laser tracker and using it to identify the model's errors, which typically improves a six-axis industrial arm's absolute accuracy several-fold to tenfold, often down to well under 1 millimeter.

ExampleA sequence of grasp points is planned offline in simulation and sent straight to the real robot: the arm returns to the same physical spot every time (good repeatability), but that spot is offset from the intended target in simulation by a consistent amount — that offset is the absolute positioning error, and in vision-guided grasping it also stacks with hand-eye calibration error.

Also called

Pose Accuracy, Absolute Accuracy

Related

Pose Repeatability · Kinematic Calibration · Offline Programming · Tool Center Point · Hand-Eye Calibration · Digital Twin

Sources

Robot calibration - Wikipedia
RoboDK Documentation: Robot Calibration (Laser Tracker)