Self-tuned closed-loop controller based on statistical data using a servomechanism

This work presents the application and use of statistical data for the design and proposal of a class of controller capable of self-tuning its gains automatically using statistical tools. The gain autotuning technique is based on sampling the output, obtaining the mean and standard deviation, the value of z in the central limit theorem. Subsequently, the probability level is estimated using the central limit theorem that allows mapping from the probability space to the profit space of the online self-tuned controller using a sigmoid type of function. Finally, some simulations and a real-time experiment with a servomechanism are presented.