Vodenje nestabilnega nelinearnega sistema z algoritmom PILCO

Mavec, Gregor (avtor), Karer, Gorazd (mentor), Kocijan, Juš (komentor)

V diplomskem delu smo ovrednotili delovanje algoritma PILCO, metode za iskanje regulacijskega zakona, na nestabilnem nelinearnem sistemu tretjega reda s kvantizirano amplitudo izhodnega signala. Na začetku so predstavljene teoretične osnove algoritma, ki zajemajo Gaussove procese in spodbujevano učenje. Sledi opis delovanja samega algoritma PILCO ter kratek primer uporabe. V tretjem poglavju je opisan matematični model sistema. Sistem, o katerem govorimo, je hidravlična naprava, kjer s črpalko spreminjamo velikost zračnega mehurčka v plovcu in tako vplivamo na njegovo globino v vodi. Vsi eksperimenti so bili izvedeni kot računalniška simulacija modela naprave. Cilj eksperimenta je voditi sistem, tako da plovec doseže in vzdržuje določeno globino. V nadaljevanju sledijo opisi potekov eksperimentov. Najprej smo, da bi ugotovili ustrezen čas vzorčenja, izvedli več eksperimentov, kjer je bila naprava vodena z regulatorjem PID. Za tem pa smo izvedli dva eksperimenta z algoritmom PILCO; prvič je bila amplituda izhodnega signala zvezna, drugič pa kvantizirana. Algoritem PILCO deluje najbolje pri vodenju sistema z zvezno amplitudo izhodnega signala. Pri sistemu s kvantizirano amplitudo izhodnega signala deluje slabše. Za iskanje uspešnega regulacijskega zakona potrebuje veliko več iteracij, referenčni vrednosti pa ne sledi več tako natančno.

• Jezik: Slovenski jezik
• Ključne besede: Gaussovi procesi, spodbujevano učenje, vodenje sistemov, računalniška simulacija
• Vrsta gradiva: Diplomsko delo
• Organizacija: FE - Fakulteta za elektrotehniko
• Leto izida: 2015
• PID: 20.500.12556/RUL-30797
• Datum objave v RUL: 19.05.2015

MAVEC, Gregor, 2015, Vodenje nestabilnega nelinearnega sistema z algoritmom PILCO [na spletu]. Diplomsko delo. [Dostopano 3 april 2025]. Pridobljeno s: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=slv&id=30797

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Control of an unstable nonlinear system using PILCO algorithm

Izvleček

In the thesis we evaluate the PILCO algorithm, a control-law search method, with a third-order unstable nonlinear system with quantized amplitude of the output signal. In the beginning the algorithm's theoretical foundation is presented, which consists of Gaussian processes and reinforcement learning. This is followed by a description of the algorithm and a short example of its operation. Next, the mathematical model of the system is described. The system is a hydraulic plant where a pump is used to influence the size of a bubble in a float, causing it to ascend or descend in the water. Instead of experiments with an actual plant, all experiments were run as computer simulations using the plant's model. The controller's objective was to stabilize the system and to reach and maintain a preselected position of the float. Then, the experimental runs with various controllers are described. Firstly, a PID controller was used for determining the appropriate sampling time of the system. This was followed by two experimental runs, where algorithm PILCO was used as a controller. First time, the amplitude of the plant's output signal was left continuous and the second time, its amplitude was quantized. Lastly, we examine the results of the experiments, and conclude that algorithm PILCO performs better with continuous signals, where it successfully meets the objective, and somewhat worse with quantized signals, where it still meets the objective but needs many more iterations.

• Ključne besede: Gaussian processes, reinforcement learning, system control, computer simulation