Automated laser beam alignment optimization using machine learning techniques
Complex light fields used in optical tweezers require advanced optical manipulation and control of the laser beam. The project focusses on the design, experimental setup and characterization of a beam auto-aligner system on a Raspberry Pi controlled stepper motor. The system will be used for maintaining and manipulating the intensity distribution of the laser beam and precise optical beamshaping by a spatial light modulator patterned optical trap for cold atoms. The work involves developing a machine learning algorithm for optimization of the “walking the beam” technique, used in most quantum optics experiments and control of structured light for advanced optical manipulation. The algorithm can be used to optimize the laser power into optical fibers, better modulation of the amplitude and phase of light and for controlling of the overlapping beams in a pump-probe experimental setup. The precise control of the laser beam intensity distribution enables the fine tuning of configurable potential wells for future optimized optical trapping experiments.