Model reference adaptive control system for laser surface treatments

This paper presents a new real time model reference neural control system for laser surface treatments, which allows to improve significantly the uniformity and final quality of the process, reducing the rejection rate and increasing the productivity and efficiency of the treatment. The proposed neural controller essentially allows to assimilate the complex system dynamics to a simple first-order linear system, which can be easily controlled by a conventional proportional, integral and derivative controller, PID, providing exceptional results.