Computer vision based positional error compensation of an industrial robot using linear regression analysis

Abstract

This paper describes the experimental procedure to determine and compensate the positional error of a five Degrees-Of-Freedoms (DOFs) industrial robot using linear regression analysis. Vision-based experiment has been conducted on an IR52C industrial robot with necessary accessories, which consists of four modules namely image interpretation, calibration, implementation and regression. Experiments have been conducted exclusively for performing regression analysis and the results of all four modules are given separately. The robot's work area is captured and the position of the target object with respect to camera's origin is determined. These camera coordinates are transformed into robot coordinates using appropriate rigid body transformations. Finally, the experimental procedure was implemented with an IR52C industrial robot and it was found that the robot is reaching the target position accurately.