DTA

Digital Theses Archive

 

Tesi etd-02292020-184443

Type of thesis
Dottorato
Author
CZIMMERMANN, TAMAS
URN
etd-02292020-184443
Title
Artificial Intelligence Strategies for Visual-Tactile Perception with Robots in Industry 4.0 and Bionics
Scientific disciplinary sector
Istituto di Biorobotica
Course
Istituto di Biorobotica - BIOROBOTICS
Committee
relatore Prof. DARIO, PAOLO
Tutor Prof. ODDO, CALOGERO MARIA
Tutor Prof. CIUTI, GASTONE
Keywords
  • Industry 4.0
  • artificial senses
  • artificial intelligence
  • collaborative robotics
  • machine vision
  • deep learning
Exam session start date
;
Availability
parziale
Abstract
Biorobotics is a new scientific technological area, that has recently evolved from the fusion of biology and robotics. It increasingly challenges robotics engineers to solve, supplement and assist scientists of biology in the quantitative demands of nature related questions. This study present 3 different biorobotics application, that share their theory to reverse engineer human behaviour, signal encoding and cognition, where the applications often maintains or even over-perform the desired standard by human needs. The author presents an industrial 4.0 platform with a multi-agent system, inspired by the human behaviour of performing a quality control task. The system successfully detected and repaired various types of defects on vehicle parts with high precision. Later, the candidate discusses a morphological characterisation strategy of a rehabilitation robotics platform approach that enabled two trans-radial amputees to successfully discriminat naturalistic textures via the MNC-based tactile feedback. As last, the author presents a novel application for collaborative robotics, that employs Fiber Bragg Grating (FBG) technology embedded into an artificial skin. The study shows, that the candidate was able to successfully localize and predict its force of the exterior mechanical interaction on a large surface of a robotic arm covered with artificial skin.
Files