Upravljanje sustavima automatiziranog staklenika pomoću IoT uređaja

master's thesis

• Title (english): Control of automated glasshouse systems by IoT Controller
• Author: Ivana Marijanović
• Mentor: Nedjeljko Perić (mentor)
• Committee member: Nedjeljko Perić (član povjerenstva)
• Committee member: Vinko Lešić (član povjerenstva)
• Committee member: Mario Vašak (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U sklopu ovog rada objašnjena je predikcija temperature automatiziranog staklenika koristeći neuronsku mrežu. Korišten je LSTM model neuronske mreže. Učenje modela je provedeno na unaprijed dobivenom skupu podataka koji je sadržavao vanjske i unutarnje uvjete („AGROSparc_measurements“). Točnost predviđanja modela procjenjuje se srednjom apsolutnom pogreškom. Testiranjem mreže različitim vrijednostima hiperparametara dobiven je konačni model koji je, kao Python skripta, učitan na Raspberry Pi. To je udaljeno mikroračunalo koje putem senzora i aktuatora komunicira sa bazom podataka. Na kraju je prikazana sinteza upravljačkog algoritma sa estimiranom vrijednošću koju daje LSTM mreža, a validacija predloženog modela na stvarnom računalu je ostavljena za buduća istraživanja.

Abstract (english)

Within this paper, it's briefly explained how to predict the temperature of an automated greenhouse using a neural network. An LSTM neural network model was used. Model learning was performed on a pre obtained set of data that maintains external and internal conditions („AGROSparc_measurements“). The accuracy of model prediction is estimated by the mean absolute error. By testing a network of different hyperparameter values, a final model was obtained, which was loaded on the Raspberry Pi. It's a remote microcomputer that communicates with databases via sensors and actuators. Finally, the synthesis of the control algorithm with the estimated value given by the LSTM network is presented, and the validation model on a real computer is set for future research.

• Keywords: model; neuronske mreže; LSTM; Python; Raspberry Pi; IoT; učenje; epohe; automatizacija; predikcija; upravljanje; temperatura; senzor; aktuator
• Keywords (english): model; neural networks; LSTM; Python; Raspberry Pi; IoT; learning; epochs; automation; prediction; control; temperature; sensor; actuator
• Language: croatian
• URN:NBN: urn:nbn:hr:168:363505
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 09:25:39