An Experimental Study of Network-Based DC Motor Speed Control using SANFIS

Network-based control (NBC) systems can provide several advantages among traditional control systems. Nevertheless, the performances of NBC systems can be degraded due to undesired network behaviors such as network-induced delays. Several NBC algorithms usually neglect several network behaviors due to assumptions in problem formulations. The incompleteness and ambiguity of this network information implies ambiguities in NBC performances. In this paper, we applied a novel NBC gain scheduling scheme by applying a SANFIS (self-adaptive neuro-fuzzy inference system) along with gain scheduling to handle ambiguities in network behaviors. The SANFIS is utilized to classify a current network condition in order to select an optimal gain for this condition. An experimental result shows that the PI controller with the proposed approach yields significantly better NBC performances.