Sustav za obradu senzorskih podataka s bicikla zasnovan na metodama strojnog učenja

master's thesis

• Title (english): System for Bicycle Sensor Data Processing Based on Machine Learning Methods
• Author: Matija Roglić
• Mentor: Mario Cifrek (mentor)
• Committee member: Mario Cifrek (predsjednik povjerenstva)
• Committee member: Hrvoje Džapo (član povjerenstva)
• Committee member: Željka Lučev Vasić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U radu je prikazan razvoj sustava na električnom biciklu koji koristi neuronsku mrežu kako bi detektirao lice korisnika i prepoznao ga. Sustav je baziran na Yocto Linux operacijskom sustavu kako bi se smanjila veličina i povećala efikasnost sustava. Na Yocto Linuxu je tijekom izgradnje dodan sloj za izgradnju TensorFlow Lite biblioteke i ROS Noetic sustava. Sam model se pokreće uz pomoć interpretera iz TensorFlow Lite biblioteke u ROS okruženju na DART-SOM-MX6 pločici, te prima ulazne podatke s prednje kamere bicikla, detektira i obrađuje lice na fotografiji te izbacuje poruku s imenom prepoznate osobe ako se ta osoba nalazi u bazi podataka.

Abstract (english)

Through the master thesis a system on an electrical bike was developed that uses neural networks to detect and recognise the user's face. The system is based on Yocto Linux operating system to decrease the size and increase the efficiency of the whole system. While building Yocto image, a layer containing recipes for building TensorFlow Lite interpreter and ROS Noetic has been added. The model is run with interpreter on a DART-SOM-MX6 board in ROS environment and it receives input from a front camera, detects the face and recognises it and sends the message about the face it recognized.

• Keywords: Neuronske mreže; ROS sustav; Yocto; Detekcija i prepoznavanje lica; električni bicikl; TensorFlow Lite
• Keywords (english): Neural networks; ROS system; Yocto; Detection and face recognition; Electrical bike; TensorFlow Lite
• Language: croatian
• URN:NBN: urn:nbn:hr:168:699044
• 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: 2023-01-13 13:43:30