Tesi etd-01202021-113040
Link copiato negli appunti
Tipo di tesi
Dottorato
Autore
AFROZ, ABANTI SHAMA
URN
etd-01202021-113040
Titolo
Human computer interface for 3D applications in industrial contexts
Settore scientifico disciplinare
ING-IND/34
Corso di studi
Istituto di Biorobotica - BIOROBOTICS
Commissione
relatore Prof. STEFANINI, CESARE
Parole chiave
- augmented reality
- CAD-based robot path planning
- human computer interface
- laser inspection
- parameter optimization for wearable device
- welding
Data inizio appello
06/04/2021;
Disponibilità
completa
Riassunto analitico
Human computer interaction (HCI) is a vast, inter-disciplinary research field involving numerous components from human anatomy and physiology, novel hardware and software interfaces, wearable sensors and actuators as well as computer vision and graphics. HCI technologies are concentrated on evoking natural human senses of vision, touch, smell, sound, and taste to provide an improved perception of objects or scenarios. A major section of this technology is dedicated towards capturing and rendering of enhanced 3D features, i.e., features that are not eminently visible to human eyes, with immersive 3D graphics technology. 3D medical diagnostic scans, industrial-grade 3D scans with sub-mm scale resolution as well as precise hand movement actions of humans are examples of some of the many enhanced information of this kind. For the remaining of the thesis, such technology, targeted towards providing improved visual perceptions, has been referred as enhanced human computer interaction (EHCI).
The concentration of this Ph.D. thesis is on adaptive processes for establishing effective human computer interfaces between a human user with objects and actions that are separated by space or time. The software and hardware components developed in this context can effectively recognize industrial context specific enhanced physical features of and relay these data to a distant person through a semi-immersive display medium in augmented reality (AR).
The first section of this thesis focuses on extracting physical features from static objects, with particular emphasis towards efficient robot programming in the context manufacturing industry. With this objective in mind, 3D inspection was performed on sample components, both in their prefabrication (computational model) and post- fabrication (real objects) phases for extracting feature parameters that are specifically useful for robot aided non-additive scenarios. Prefabrication analyses were done on finished CAD models represented in .STL file formats. Post fabrication analyses were performed on data collected for the same samples but obtained with reverse engineering technologies. Comparisons were also performed between results obtained in these two stages. The innovative methodologies applied in these phases can greatly contribute in the field of manufacturing process planning and industrial in-process inspections.
The thesis also focuses on capturing refined dynamic motion in 3D space with particular interests on refined human hand interactions and with this motivation an optimization system was developed for improving the performance of an opto-inertial hand motion tracking prototype that has been developed by the author’s research group. The obtained advancements can be exploited in the fields of hand motion mimicking robot programming as well as manual hand motion and performance analysis, both in industrial (e.g., hand-held device’s speed quantification) and clinical (including rehabilitation performance evaluation) contexts.
A further contribution to this thesis is relaying these enhanced data to a spatiotemporally distant human user. This was achieved by setting up a customized semi-immersive virtual environment. An OpenCV 4.5 based system was developed in C++ language that can utilize on and off line workspace vision of a human user to convey enhanced data, that is separated by space or time or both. For this thesis the system has been restricted to display only static data while offering the possibility to enhance it’s capacity to display time variant data as well.
Enhanced human computer interactions offer immense opportunities including and not limited to industrial, clinical and ‘remote working’ contexts. This doctoral dissertation describes, in its different case studies, the perceived approaches to effectively establish human-computer interactions for inspecting and experiencing remote information by bridging human centric designs with advanced artificial monitoring systems and aims to contribute in paving the way for real world applications of immersive human computer interfacing technologies.
The concentration of this Ph.D. thesis is on adaptive processes for establishing effective human computer interfaces between a human user with objects and actions that are separated by space or time. The software and hardware components developed in this context can effectively recognize industrial context specific enhanced physical features of and relay these data to a distant person through a semi-immersive display medium in augmented reality (AR).
The first section of this thesis focuses on extracting physical features from static objects, with particular emphasis towards efficient robot programming in the context manufacturing industry. With this objective in mind, 3D inspection was performed on sample components, both in their prefabrication (computational model) and post- fabrication (real objects) phases for extracting feature parameters that are specifically useful for robot aided non-additive scenarios. Prefabrication analyses were done on finished CAD models represented in .STL file formats. Post fabrication analyses were performed on data collected for the same samples but obtained with reverse engineering technologies. Comparisons were also performed between results obtained in these two stages. The innovative methodologies applied in these phases can greatly contribute in the field of manufacturing process planning and industrial in-process inspections.
The thesis also focuses on capturing refined dynamic motion in 3D space with particular interests on refined human hand interactions and with this motivation an optimization system was developed for improving the performance of an opto-inertial hand motion tracking prototype that has been developed by the author’s research group. The obtained advancements can be exploited in the fields of hand motion mimicking robot programming as well as manual hand motion and performance analysis, both in industrial (e.g., hand-held device’s speed quantification) and clinical (including rehabilitation performance evaluation) contexts.
A further contribution to this thesis is relaying these enhanced data to a spatiotemporally distant human user. This was achieved by setting up a customized semi-immersive virtual environment. An OpenCV 4.5 based system was developed in C++ language that can utilize on and off line workspace vision of a human user to convey enhanced data, that is separated by space or time or both. For this thesis the system has been restricted to display only static data while offering the possibility to enhance it’s capacity to display time variant data as well.
Enhanced human computer interactions offer immense opportunities including and not limited to industrial, clinical and ‘remote working’ contexts. This doctoral dissertation describes, in its different case studies, the perceived approaches to effectively establish human-computer interactions for inspecting and experiencing remote information by bridging human centric designs with advanced artificial monitoring systems and aims to contribute in paving the way for real world applications of immersive human computer interfacing technologies.
File
| Nome file | Dimensione |
|---|---|
| ThesisManuscript.pdf | 6.96 Mb |
Contatta l'autore |
|