Sustav za stvarnovremensku procjenu parametara vožnje bicikla temeljen na metodama strojnog učenja

master's thesis

• Title (english): Real-Time System for Bicycle Riding Parameters Estimation Based on Machine Learing Methods
• Author: Andrija Mijić
• Mentor: Hrvoje Džapo (mentor)
• Committee member: Hrvoje Džapo (predsjednik povjerenstva)
• Committee member: Mario Cifrek (član povjerenstva)
• Committee member: Željka Lučev Vasić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-02-25, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U okviru diplomskog rada razvijen je sustav za prikupljanje podataka na električnom biciklu. Sustav se sastoji od STM32F4 Discovery kao centralne procesorske jedinice, inercijskog senzora ICM-20948, barometra DPS368 Shield2Go i RTK sustava za pozicioniranje Emlid Reach. Senzorski podaci snimaju se na SD karticu, pri čemu je posebna pažnja posvećena vremenskoj sinkronizaciji podataka iz različitih izvora. Podaci dobiveni eksperimentalnom vožnjom obrađeni su i korišteni za treniranje umjetne neuronske mreže. Za treniranje mreža korišteni su alati Tensorflow i Google Colab, pri čemu je Tensorflow korišten za razvoj umjetne neuronske mreže dok je Google Colab korišten kao razvojno okruženje. Proučene su mogućnosti postavljanje trenirane umjetne neuronske mreže na ugradbeni računalni sustav DART-MX.

Abstract (english)

In this thesis a system for sensor data acquisition for electric bicycle was developed. The system consists of: STM32F4 Discovery board (CPU unit), ICM-20948 inertial sensor, DPS368 Shield2Go barometer, and Emlid Reach RTK positioning system. Sensor data was recorded on the SD card, with special attention to the time synchronization of sensor data obtained from different sources. Data obtained by experimental driving were processed and used for training of artificial neural network. Tensorflow and Google Colab were used to train neural networks, where Tensorflow was used to develop neural network while Google Colab was used as a development environment. Process of transfering a neural network on the embedded computer system DART-MX was studied.

• Keywords: ugradbeni računalni sustavi; inercijski senzori; barometar; Real-time Kinematics (RTK); neuronske mreže; Tensorflow
• Keywords (english): embedded systems; inertial sensors; barometer; Real-time Kinematics (RTK); neural networks; Tensorflow
• Language: croatian
• URN:NBN: urn:nbn:hr:168:584821
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 13:07:50