Nonlinear Controller for the Laser Fiber Using PID Controller

Abstract

We show that a software architecture based on machine learning and adaptive control advances makes self-tuning optics possible. Commercially available optical telecom components may be combined with servo controllers to develop a training and execution software module that can self-tune the laser cavity even when it is switched off. Mechanical and/or environmental disturbances may aid in frequency comb stabilization. An exhaustive search of state space is used in the algorithm training stage to find the optimum performance areas for one or more objectively interesting functions. The technique of implementation stage starts by identifying the variable space using a sparse sensing approach, then settles on a near-optimal solution and maintains it using the extremum seeking control protocol.