Asynchronous bio-inspired tuning for the DC motor speed controller with simultaneous identification and predictive strategies

One of the main issues in the control system is the online tuning of its gains. The use of bio-inspired algorithms (BA) is gaining more attention in the control tuning task because they are less sensible to system uncertainties. Nevertheless, the computational time of BA must be reduced to be used in practice. In this work, an event condition is stated to reduce the computational cost of the optimization process in the online bio-inspired tuning approach. This condition activates the tuning approach only when it is required, i.e., when the regulation error tends to increase. Also, in this approach, an identification process and a predictive strategy are simultaneously optimized to find the more suitable control parameters that handle more efficient the parametric uncertainties. The proposed online Asynchronous Bio-inspired Tuning Approach with Simultaneous Identification and Prediction (ABioTASIP) is validated in the study case of the velocity regulation of a DC motor considering dynamic parametric uncertainties. The comparative analysis with an approach where the control parameters are periodically tuned indicates that the proposal decreases the tuning process without considerably increase the regulation error.