Högskolan i Skövde

The industrial use of Virtual Reality, Augmented Reality, Mixed Reality and all the eXpanded Realities is rapidly increasing over the last few years. This thesis aims to provide with a standardised method to measure calibration alignment error for augmented reality devices. A literature review was conducted on different SLAM and calibration methods to achieve this goal. The device selected for this thesis was HoloLens2, as it has a considerably better performance than its competitors. For the experiment, the first step needed was to create a digital replica of a small robot on RobotStudio.This study proposes a method to measure alignment calibration error based on projecting the hologram of a cube against a physical calibrated board to measure error on each axis. The deviation of the hologram from its calibrated position can be calculated directly setting the hologram to locate it at the 0,0,0 of the real calibrated board. Validation is done by doing only one calibration, using several holograms set at known distances which will cover a wide range of positions. For the validation to be successful, the initial condition is that the orientation error needs to be close to cero or relatively negligible in comparison with the linear alignment error, as the proposed method it is not able to measure orientation error and alignment error simultaneously. If validation is successful, the calibration alignment error of any calibration method can be measured with this method by projecting a hologram against the calibrated board and repeating the experiment several times to perform a statistical analysis of the error measured. Finally, this study suggests some calibration methods which may give better results than the one evaluated. It discusses that automatic calibration methods will very likely be the future for calibration. Still, there is a precise manual method which may be an excellent fit to calibrate the devices when used to display safety areas, which is also suggested. As the experiment could not be carried out due to several limitations, there are not results of the calibration alignment error. However, this thesis gives a detailed guideline on this thesis to be able to replicate the experiment at any time for future work.