An approach to human behavior adaptable collaborative robot workplace design

Produktbeschreibung

Studies show that collaborative robots that can explain in real time what a human is doing and the passive underlying intention are likely to be positively perceived by human workers. In this respect, a seamless flow of information between human and machine systems in realtime is expected. This will push future collaborative robots‘ ability to work with humans at higher collaboration levels, including real cooperation or responsive collaboration with humans. This is crucial for collaborative robot applications to be used not in a similar way traditional industrial robots are used, i.e., rather than replacing humans engaging in real collaboration with them. Therefore, this work presents a bottom-up method for collaborative robot workplace design that uses models of individual human attention, intention, and action for cycle time synchronization to improve the performance of action prediction in planning the collaborative robot workplace design process.

Contents
Abbreviations and Acronyms VIII
Symbols X
Abstract XI
Zusammenfassung XII
1 Introduction 1
1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Human individual collaborative robot workplace . . . . . . . . . . . . . . . 7
1.2.1 Collaborative robot workplace design principle . . . . . . . . . . . . 7
1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5 Structure and organization . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2 State of the art 17
2.1 Background of workplace design methods . . . . . . . . . . . . . . . . . . . 17
2.1.1 Top down collaborative robot workplace design approach . . . . . . 20
2.1.2 Bottom up collaborative robot workplace design approach . . . . . 21
2.1.3 Outside-in collaborative robot workplace design approach . . . . . . 24
2.2 Human motion behavior capturing for collaborative workplace design . . . 26
2.2.1 Digital human modeling . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2.2 Digital collaborative robot modeling . . . . . . . . . . . . . . . . . 33
2.2.3 Workplace components design and model . . . . . . . . . . . . . . . 34
2.3 Human action, attention and intention modeling . . . . . . . . . . . . . . . 37
2.3.1 Motion time based human activity models . . . . . . . . . . . . . . 37
2.3.2 Task-based building-block human activity models . . . . . . . . . . 37
2.3.3 Space-time based human activity models . . . . . . . . . . . . . . . 38
2.3.4 Stochastic based human activity models . . . . . . . . . . . . . . . 39
2.3.5 Rules-based human activity models . . . . . . . . . . . . . . . . . . 40
2.3.6 Anticipatory human action model . . . . . . . . . . . . . . . . . . . 41
3 Individualized collaborative robot workplace design (iCRoWD) 44
3.1 iCRoWD method description . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.2 Attention-Intention-Action (AIA) model . . . . . . . . . . . . . . . . . . . 50
3.2.1 Concept of AIA formulation . . . . . . . . . . . . . . . . . . . . . . 50
3.2.2 AIA modeling and mapping . . . . . . . . . . . . . . . . . . . . . . 55
3.2.3 Temporal integration . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.4 Adaptability analysis – capability, reach and collision . . . . . . . . 67
4 Design of experiment and implementation 69
4.1 Physical and digital workplace setup . . . . . . . . . . . . . . . . . . . . . 69
4.2 Motion capturing system set-up and interface configuration . . . . . . . . . 72
4.2.1 Motion capturing system set-up . . . . . . . . . . . . . . . . . . . . 72
4.2.2 Motion capture systems interface . . . . . . . . . . . . . . . . . . . 74
4.3 iCRoWD: Human motion behavior analysis and characterization . . . . . . 76
4.3.1 iCRoWD approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3.2 Simulation approach . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.4 Comparison and evaluation of the iCRoWD approach . . . . . . . . . . . . 79
4.4.1 Measures definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.4.2 KPI definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5 Tests and results 87
5.1 Test scenario overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2 Scenario I: Pick and place operation for sorting parts in lab environment . 88
5.3 Scenario II: Fixture assembly in lab environment . . . . . . . . . . . . . . . 94
5.4 Scenario III: Industrial automotive gas tight hose assembly (nonexperienced worker) . . . . . . . . . . . . 101
5.5 Scenario IV: Industrial automotive gas tight hose assembly (experienced
worker) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.6 Action duration evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.7 Adaptability analysis for collaborative robot workplace based on result
summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.8 Result validation and verification . . . . . . . . . . . . . . . . . . . . . . . 121
6 Discussions 127
6.1 Result interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
VIContents
6.2 Individualized human behavior analysis . . . . . . . . . . . . . . . . . . . . 128
6.3 Collaborative robot workplace adaptation . . . . . . . . . . . . . . . . . . 133
6.4 Potentials and limitations of iCRoWD . . . . . . . . . . . . . . . . . . . . 135
7 Conclusion and outlook 137
7.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.2 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
A Appendices 141
A.1 Terminologies in human action identification . . . . . . . . . . . . . . . . . 141
A.2 Digital twin for modeling for collaborative robot workplaces . . . . . . . . 145
A.3 Systematic literature analysis for human action modeling and simulation . 153
A.4 Extended test and verification of the iCRoWD method . . . . . . . . . . . 159
A.5 Gaze analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
A.6 HAROPP: Human activity recognition based on probabilistic partition . . 162
A.7 MOSIM approach for human motion simulation . . . . . . . . . . . . . . . 162
A.8 HARNets: Deep learning based memory networks for human action recognition . . . . . . . . . . . 167
A.9 Relevant standards and lists of some industrial collaborative robots . . . . 176
Bibliography 179

Keywords: Aufmerksamkeit-Intention-Handlung, menschliches Handlungsmodell, Mensch-RoboterKollaboration, kollaborative Roboterarbeitsplatzgestaltung, symbiotischer Arbeitsplatz, Attention-intention-action, human action model, human-robot collaboration, collaborative robot workplace design, symbiotic workplace