Hand-eye calibration in the context of Augmented Reality

Abstract: The calibration between a camera and an industrial robot is a well-established area of research. Nevertheless, traditional setups concentrate on the case in which the camera is either attached to the robot or static. With the advent of Augmented Reality (AR) capable devices, such as headsets, glasses, phones, or tablets, the device's camera is now free to roam the scene in an unconstrained manner.In order to unlock a new set of applications, the AR device needs to be aware of the precise location of the robot. Unfortunately, this new embodiment of robot to moving camera calibration has comparatively received much less attention. In this thesis, we will explore the aforementioned setup motivated by a real use case at ABB robotics. After exploring the mathematical preliminaries, we will analyze the problem from two different mathematical formalisms and implement the solution in a resource-constrained device, leading to a new patent pending approach.