20.500.12556/DKUM-84946 Samosprožilno mrežno vodenje z nelinearnim modelom na osnovi globokega učenja magistrsko delo Self-triggered network control with a nonlinear model based on deep learning Magistrska naloga opisuje modeliranje nelinearnih dinamičnih sistemov in implementacijo samosprožilnega dogodkovnega vodenja na sistemu zračne levitacije z namenom reševanja sodobnih problemov vodenja, kot je preobremenjenost omrežja. Najprej smo vzpostavili komunikacijo med sistemom in računalnikom z namenom priprave podatkov. Sledila je faza globokega učenja in validacija modela. Na koncu smo načrtali ustrezen algoritem, ki posodablja izhod regulatorja glede na predikcijo modela. Z magistrskim delom smo predvsem dokazali delovanje obravnavanega vodenja na hitrem nelinearnem in nestabilnem sistemu. Ugotovili smo, da je zanesljivost takega vodenja predvsem odvisna od natančnosti modela. Samosprožilno vodenje je lahko riskantno, zato je za industrijsko aplikacijo potrebno vpeljati dodatne varnostne mehanizme. The master's thesis describes the modelling of nonlinear dynamical systems and the implementation of a self-triggered control on the air levitation system with the aim of solving modern control problems such as network overload. First, we established communication between the system and the computer for the purpose of data preparation. This is followed by the phase of deep learning and model validation. Finally, we designed an appropriate algorithm that updates the controller output according to the model prediction. With the master's thesis, we mainly proved the operation of the discussed control on a fast nonlinear and unstable system. We found that the reliability of this control depends mainly on the accuracy of the model. Self-triggered control can be risky and therefore it is necessary to introduce additional safety mechanisms for an industrial application. samosprožilno vodenje dogodkovno proženje nelinearni model globoko učenje NARX. self-triggered control event based sampling nonlinear model deep learning NARX. true true false [S. Vogrinčič] Slovenski jezik Angleški jezik Magistrsko delo/naloga 2023-08-10 10:24:56 2023-10-05 16:16:01 2023-10-27 03:12:24 0000-00-00 00:00:00 2023 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF (XII, 61 f.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 681.521:004.85(043.2) 170018563 MAG_Vogrincic_Sebastjan_2023.pdf MAG_Vogrincic_Sebastjan_2023.pdf 1 639EF1AE5C962455EC25AF8FEA00B610 8230b991633da47e3fc3e7993c236c5f9f463f52abcd37700678fd0c92c4c69b 60e64982-465a-11ee-93d5-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=170684 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0