Podatna manipulacija osjetljivih predmeta paralelnom hvataljkom

master's thesis

• Title (english): Compliant parallel gripper manipulation for soft objects
• Author: Jelena Vuletić
• Mentor: Matko Orsag (mentor)
• Committee member: Matko Orsag (predsjednik povjerenstva)
• Committee member: Stjepan Bogdan (član povjerenstva)
• Committee member: Zdenko Kovačić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Rad se bavi problemom izgradnje 3D modela osjetljivih objekata na osnovu podataka sa RGB-D kamere i senzora sile koji su pričvršćeni na Franka Emika Panda robota. Sustav za podatnu manipulaciju objekata, čiji je izlaz 3D model, razvijen je u simulacijskom okruženju i sastoji se od nekoliko dijelova. Prvi dio sustava je detekcija objekata. Problem detekcije riješen je korištenje neuronske mreže. Nakon detekcije, prelazi se na planiranje gibanja koje je ostvareno korištenjem MoveIt!-a. Problem izbora točke hvata također je riješen korištenjem neuronske mreže. Za upravljanje paralelnom hvataljkom robotskog manipulatora koristi se impedantni regulator. Konačno, za izradu 3D modela korišten je point cloud zapis. U sklopu rada prikazana je teorijska podloga svih navedenih podzadataka te je razvijen je ROS paket koji omogućuje pokretanje potpuno automatiziranog sustava za podatnu manipulaciju objekata.

Abstract (english)

The thesis deals with the problem of creating a 3D model of soft objects based on the data from RGB-D camera and force sensors attached to Franka Emika Panda robot. System for compliant manipulation, which outputs such 3D model, is developed in simulation environment and it consists of several parts. First part of the system is an object detection. Object detection problem is solved using neural network. Detection is followed by motion planning, which is achieved using MoveIt! framework. Grasping problem is solved using neural network approach. Impedance control is implemented for controlling a parallel gripper. Finally, 3D model is written in the point cloud format. Theoretical background of all of the sub-tasks is given in the thesis alongside with the implementation of developed algorithms, which can be run completely autonomously using ROS package which was developed as a part of the thesis.

• Keywords: Detekcija objekata; Planiranje gibanja; Izbor točke hvata; Impedantno upravljanje
• Keywords (english): Object detection; Motion planning; Grasping; Impedance control
• Language: croatian
• URN:NBN: urn:nbn:hr:168:799762
• 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: 2023-01-15 11:22:01