NN-ANARX model based control of liquid level using visual feedback
Kristina Vassiljeva, Aleksei Tepljakov, Eduard Petlenkov · 2015
In this paper, the problem of liquid level control based on visual feedback is investigated in application to a real life model of an industrial plant. Visual detection of liquid level is implemented on the Raspberry Pi computer. Computational intelligence based controller uses input-output feedback linearization. Parameters of the controller are provided by the neural network ANARX structure. As communication between Raspberry Pi and control system is provided over WLAN, additional prediction module is used to overcome networking problems. Control of the SISO and MIMO systems is provided.