Development of Techniques for a Straight-Line Alignment System Based on Pseudo-Nondiffracting Structured and Layer Laser Beams

This thesis develops solutions based on pseudo-nondiffracting Structured Laser Beams (SLBs) and Layer Beams (LBs) for alignment systems based on a straight-line reference. The thesis is organized into a theoretical part, followed by six thematically interconnected publications. The first publication describes the generating wavefront reconstruction method using a convolutional neural network and least-squares fitting to estimate defocus and spherical aberrations from a single transversal intensity distribution, which achieved the root-mean-square error (RMSE) of the reconstruction 0.09λ relative to Shack–Hartmann sensor measurements. The second publication investigates how atmospheric effects influence the quality of the alignment reference. The third publication build on this work by presenting, together with the second, an analysis of centroiding algorithms applied to SLBs and Hollow SLBs. On simulated datasets, the algorithms introduced absolute detection errors in the range of 10−4 pixel (px) to 10−2 px, while the experimental validation of precision demonstrated a standard deviation of 1.3 μm. The fourth publication describes experimental generation of LBs up to 140 m and presents an algorithm for alignment reference plane detection, which introduced a standard deviatio error equal to 0.7 μm for the experimental validation of precision.The fifth publication describes the generation principle of LBs in detail. The last publication presents an LB-based alignment sys-tem prototype, comparing it to state-of-the-art measurements. The results showed a repeatability and reproducibility of the alignment reference with an RMSE below 30.0 μm. The relative alignment precision was validated as 4.3 μm over a 2 m distance. The last paper also describes a conceptual idea for a fiber matrix-based radiation-hard sensor. The proof-of-concept yielded a root mean squared error of the detection of the reference below 1.3 μm. This error was induced by the fiber matrix. The results highlight the potential for using straight-line reference systems based on SLBs and LBs for the alignment of particle accelerators at CERN.