Nonlinear controller for the laser fiber using PID controller

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