Integracija robotskih senzorskih sustava za interakciju sa složenim površinama

master's thesis

• Title (english): Integration of robotic sensory systems for interaction with complex surfaces
• Author: Albert Androšić
• Mentor: Filip Šuligoj (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2024-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering; Precision Mechanical Engineering

Abstract

Cilj ovog diplomskog rada je integracija robota Franke Emike u ROS2 operativni sustav i njegovo osposobljavanje za praćenje zakrivljene 3D površine pomoću ugrađenih senzora sile i momenta. U radu je naveden postupak instalacije ROS2 operativnog sustava kao i pripadajućih podsustava za uspješno spajanje robota s računalom. Detaljno su opisani postupci integracije robota Franka Emika u ROS2 okruženje, prilagodba postojećih ROS2 paketa poput MoveIt2 te njihova prilagodba potrebama ovog istraživanja. Za snimanje radne okoline korištena je Intel RealSense D435i kamera, a njezine fizičke karakteristike, zajedno s instalacijom potrebnih biblioteka i alata za upotrebu u ROS2 sustavu te postupak „eye-in-hand“ kalibracije za što precizniju vizualizaciju prostora detaljno su opisani. Navedena „eye-in-hand“ kamera upotrijebljena je za generiranje 3D slike pomoću oblaka točaka pomoću kojeg korisnik u ROS2 okruženju odabire pozicijske markere kao ciljne točke robotske ruke. Što se konstrukcijskog aspekta ovog rada tiče, u radu su dokumentirani modeli nosača dubinske kamere za „eye-in-hand“ konfiguraciju i ticala sa sfernim vrhom, koji se može montirati na prihvatnicu i uhvatiti pomoću pomičnih prstiju. Modeli su konstruirani pomoću 3D pisača i montirani na prirubnicu, odnosno prihvatnicu. Radni komad korišten za testiranje ovog sustava nije dizajniran za potrebe ovog rada, već je to testni uzorak laboratorija CRTA-e koji se koristi u nekoliko projekata.
Temelj ovog rada leži u izradi kontrolera koji omogućuje precizno praćenje zakrivljene površine te održavanje konstantnog kontakta između radnog alata i površine regulacijom sile i održavanjem njezine vrijednosti unutar zadanih granica. Sedam momentnih senzora raspoređenih u svim zglobovima robota koriste se za estimaciju sile u radnom alatu, koja se potom regulira putem diskretnog pomaka robota duž zadane putanje. Putanja od točke do točke definirana je pomoću MoveIt2 paketa s ciljem optimizacije procesa. Provedena je eksperimentalna validacija točnosti pozicioniranja radnog alata korištenjem kalibracijske ploče, koja se koristi i za kalibraciju kamere te validacija praćenja 3D zakrivljene površine pomoću diskretnog kontrolera sile i kombinacije triju ili četiriju točaka.

Abstract (english)

The goal of this master’s thesis is the integration of Franka Emika robot in the ROS2 operating system and its adaptation for tracking a curved 3D surface tracking using built-in force and torque sensors. The thesis outlines the process of installing ROS2 operating system and its subsystems for successfully connecting the robot to the computer. The procedures for integrating Franka Emika robot into the ROS2 environment, adapting existing ROS2 packages such as MoveIt2, and customizing them to suit the needs of this research are also described in detail.
For capturing the working environment, an Intel RealSense D435i camera was utilized. Its physical characteristics, along with the installation of necessary libraries and tools for use in the ROS2 system, as well as the “eye-in-hand” calibration process for precise spatial visualization are thoroughly explained. The mentioned “eye-in-hand” camera was used to generate a 3D image through point cloud data, allowing the user in the ROS2 environment to select positional markers as target points for the robotic arm. Regarding the construction aspect of this paper, the thesis documents the models of the depth camera mount for the "eye-in-hand" configuration and the tactile tool with a spherical tip, which can be mounted on a robot gripper and gripped by movable fingers. These models were constructed using a 3D printer and mounted on the flange and the gripper. The workpiece used for testing this system was not designed specifically for this thesis but is a sample from the CRTA laboratory used in several projects.
The foundation of this research lies in the development of a controller enabling precise tracking of the curved surface and maintaining constant contact between the tool and the surface by regulating force and keeping its value within specified limits. Seven torque sensors distributed across all robot joints are employed to estimate the force at the tool, which is then regulated by discrete robot movements along the defined path. The point-to-point path is defined using MoveIt2 package to optimize the process. An experimental validation of the accuracy of tool positioning was conducted using a calibration board, which is also used for camera calibration. Also, the tracking of the 3D curved surface using the discrete force controller was validated using four and three point combinations.

• Keywords: Franka Emika; ROS2; oblak točaka; MoveIt2; konstantno održavanje sile; praćenje oblika površine
• Keywords (english): Franka Emika robot; Pointcloud; ROS2; MoveIt2; constant force maintenance; following the shape of the surface
• Language: croatian
• URN:NBN: urn:nbn:hr:235:046376
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-01-22 18:15:32