Trends in Control and Decision-Making for Human–Robot Collaboration Systems (eBook)

Yue Wang received her B.S. degree in Mechanical Engineering from Shanghai University, China, in 2005 and M.S. and Ph.D. degrees in Mechanical Engineering from Worcester Polytechnic Institute in 2008 and 2011. She is an Assistant Professor in the Department of Mechanical Engineering at Clemson University. Prior to joining Clemson in 2012, she was a postdoctoral research associate in the Electrical Engineering Department at the University of Notre Dame. Her research interests include cooperative control and decision-making for human-robot collaboration systems, multi-agent systems, and control of cyber-physical systems. Dr. Wang received the National Science Foundation CAREER award and the Air Force Summer Faculty Fellowship in 2015, respectively. Her research has lead to 9 journal publications, 23 peer-reviewed conference papers, a book, and 2 book chapters. Dr. Wang is a member of IEEE, ASME, and AIAA. She is the co-chair for the IEEE Technical Committee on Manufacturing Automation and Robotic Control and organizers of several invited sessions in the American Control Conference. Fumin Zhang is Associate Professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology. He received a PhD degree in 2004 from the University of Maryland (College Park) in Electrical Engineering, and held a postdoctoral position in Princeton University from 2004 to 2007. His research interests include mobile sensor networks, maritime robotics, control systems, and theoretical foundations for cyber-physical systems. He received the NSF CAREER Award in September 2009, the Lockheed Inspirational Young Faculty Award in March 2010, the ONR Young Investigator Program Award in April 2010, and the GT Roger P. Webb Outstanding Junior Faculty Award in April 2011. He is currently serving as the co-chair for the IEEE RAS Technical Committee on Marine Robotics, and the chair for the IEEE CSS Technical Committee on Robotic Control and Manufacturing Automation.

Preface 6
Acknowledgements 7
Contents 8
1 Introduction 17
1.1 Overview 17
1.2 Collaboration Between One Human--Robot Pair 20
1.3 Collaboration Between Human and Multiple Robots/Swarms 22
References 24
2 Robust Shared-Control for Rear-Wheel Drive Cars 30
2.1 Introduction 30
2.2 Problem Formulation, Definitions, and Assumptions 31
2.3 Design of the Shared-Control Law with Measurements of Absolute Positions 34
2.3.1 Design of the Feedback Controller 34
2.3.2 Shared-Control Algorithm 37
2.4 Disturbance Rejections 40
2.5 Design of the Shared Control Without Measurements of Absolute Positions 43
2.5.1 Design of the Feedback Controller 44
2.5.2 Shared-Control Algorithm 46
2.6 Case Studies 48
2.6.1 Case I: Turning Without Absolute Positioning 48
2.6.2 Case II: Driving on a Road with Parked Cars 51
2.6.3 Case III: Emergency Breaking 51
2.7 Conclusions 53
References 53
3 Baxter-On-Wheels (BOW): An Assistive Mobile Manipulator for Mobility Impaired Individuals 56
3.1 Introduction 56
3.2 System Description 59
3.2.1 Experimental Platform: BOW 59
3.2.2 System Kinematics 63
3.3 Control Algorithm 64
3.3.1 Baseline Shared-Control Algorithm 64
3.3.2 Free-Space Mode and Contact Mode 66
3.4 Application to the BOW 69
3.4.1 User Interface 69
3.4.2 Object Pick-Up and Placement Task 70
3.4.3 Board Cleaning Task 73
3.5 Conclusion 75
References 76
4 Switchings Between Trajectory Tracking and Force Minimization in Human--Robot Collaboration 79
4.1 Introduction 79
4.2 Dynamic Models 81
4.2.1 Robot Model 82
4.2.2 Human Arm Model 82
4.2.3 Unified Model 84
4.2.4 Trajectory Tracking 85
4.3 Control Design 85
4.3.1 Control Objective 85
4.3.2 Selection of Cost Functions 86
4.3.3 Optimal Control 87
4.4 Simulations 89
4.4.1 Simulation Settings 89
4.4.2 Change of Weights 90
4.4.3 Adaptation of Desired Trajectory 92
4.5 Conclusions 93
References 94
5 Estimating Human Intention During a Human--Robot Cooperative Task Based on the Internal Force Model 96
5.1 Introduction 96
5.2 Internal Force Model 99
5.2.1 Problem Formulation 99
5.2.2 Existing Models 100
5.2.3 Proposed Model 101
5.2.4 Discussion 104
5.3 Method 104
5.3.1 Apparatus 105
5.3.2 Procedure 105
5.4 Results 107
5.5 Validation of the Model 110
5.6 Statistical Analysis of the Internal Force Features 112
5.6.1 Initial Grasp Force Magnitude 113
5.6.2 Final Grasp Force Magnitude 114
5.6.3 Internal Force Energy 114
5.6.4 Difference Between Initial and Final Grasp Forces 115
5.6.5 Internal Force Variation 115
5.6.6 Negotiation Force 116
5.6.7 Negotiation Force Versus Object Velocity 117
5.7 Proposed Cooperation Policy 118
5.8 Conclusion 120
References 121
6 A Learning Algorithm to Select Consistent Reactions to Human Movements 123
6.1 Introduction 123
6.2 Background 125
6.2.1 Expert-Based Learning 125
6.2.2 Binary Learning Algorithms 126
6.3 Analysis 127
6.3.1 Performance 128
6.3.2 Consistency 128
6.3.3 Adaptiveness 130
6.3.4 Tie Breaking 131
6.4 Expanded Dual Expert Algorithm 132
6.4.1 Performance Analysis 133
6.4.2 Consistency and Adaptiveness 134
6.5 Simulation 134
6.5.1 Dual Expert Algorithm 134
6.5.2 Expanded Dual Expert Algorithm 135
6.6 Experiment 137
6.6.1 Setup 138
6.6.2 Results 139
6.7 Conclusions 141
References 141
7 Assistive Optimal Control-on-Request with Application in Standing Balance Therapy and Reinforcement 143
7.1 Introduction 143
7.2 Assistive Control Synthesis 145
7.2.1 Calculating a Schedule of Optimal Infinitesimal Actions 145
7.2.2 Computing the Control Duration 150
7.3 Human--Robot Interaction in Assisted Balance Therapy 151
7.3.1 Related Work: Assist-as-Needed Techniques 152
7.3.2 Interactive Simulation Study 153
7.4 Human--Robot Communication in Posture Reinforcement: A Short Study 157
7.5 Conclusion 160
References 161
8 Intelligent Human--Robot Interaction Systems Using Reinforcement Learning and Neural Networks 164
8.1 Introduction 164
8.2 HRI Control: Motivation and Structure Overview of the Proposed Approach 166
8.3 Inner Robot-Specific Loop 167
8.4 Outer Task-Specific Loop Control 173
8.4.1 Task-Specific Outer Loop Control Method: An LQR Approach 173
8.4.2 Learning Optimal Parameters of the Prescribed Impedance Model Using Integral Reinforcement Learning 177
8.5 Simulation Results 178
8.6 Conclusion 185
References 185
9 Regret-Based Allocation of Autonomy in Shared Visual Detection for Human--Robot Collaborative Assembly in Manufacturing 188
9.1 Introduction 188
9.2 The Hybrid Cell for Human--Robot Collaborative Assembly 190
9.3 Detection Problem Formulation with Focus on the Selected Assembly Task 193
9.3.1 Description of the Problem 193
9.3.2 Problem Formulation 196
9.4 Bayesian Sequential Decision-Making Algorithm for Allocation of Autonomy 197
9.5 Inclusion of Regret in Bayesian Decision-Making Algorithm for Allocation of Autonomy 198
9.6 Illustration of the Decision-Making Approach 201
9.6.1 Illustration of the Optimal Bayesian Decision-Making Approach 201
9.6.2 Illustration of the Regret-Based Modified Decision-Making Approach 204
9.7 Implementation Scheme of the Regret-Based Bayesian Decision-Making Approach for the Assembly Task 204
9.7.1 The Overall Scheme in a Flowchart 204
9.7.2 Measurement of Sensing Probability and Observation Cost 207
9.7.3 Measurement Method for Regret Intensity 208
9.8 Experimental Evaluation of the Regret-Based Bayesian Decision-Making Approach 210
9.8.1 Objective 210
9.8.2 Hypothesis 210
9.8.3 The Evaluation Criteria 211
9.8.4 The Experiment Design 211
9.8.5 Subjects 211
9.8.6 The Experimental Procedures 212
9.9 Evaluation Results and Analyses 212
9.10 Conclusions and Future Innovations 214
References 215
10 Considering Human Behavior Uncertainty and Disagreements in Human--Robot Cooperative Manipulation 217
10.1 Introduction 217
10.2 Human--Robot Cooperative Manipulation 219
10.2.1 Cooperative Manipulation 219
10.2.2 Control Challenges in Physical Human--Robot Interaction 222
10.2.3 Reactive Assistants 222
10.2.4 Proactive Assistants 223
10.3 Interaction Wrench Decomposition 225
10.3.1 Nonuniform Wrench Decomposition Matrices 226
10.3.2 Effective and Internal Wrenches 227
10.3.3 Load Share and Disagreement 231
10.4 Optimal Robot Assistance Considering Human Behavior Uncertainty and Disagreements 231
10.4.1 Anticipatory Assistance Based on Learned Models 232
10.4.2 The Two-Dimensional Translational Case 237
10.4.3 Experiments 239
10.5 Conclusions 245
References 248
11 Designing the Robot Behavior for Safe Human--Robot Interactions 251
11.1 Introduction 251
11.1.1 The Safety Issues and Existing Solutions 252
11.1.2 Safety Problems in HRI: Conflicts in Multiagent Systems 252
11.1.3 Safe Control and Exploration 253
11.2 Modeling the Human--Robot Interactions 254
11.2.1 The Agent Model 254
11.2.2 The Closed-Loop System 255
11.2.3 Information Structure 256
11.3 The Safety-Oriented Behavior Design 257
11.3.1 The Safety Principle 257
11.3.2 The Safety Index 258
11.4 The Safe Set Algorithm (SSA) 260
11.4.1 The Control Algorithm 261
11.4.2 Online Learning and Prediction of Humans' Dynamics 262
11.4.3 Applications 263
11.5 The Safe Exploration Algorithm (SEA) 265
11.5.1 The Safe Set in the Belief Space 266
11.5.2 Learning in the Belief Space 267
11.5.3 A Comparative Study Between SSA and SEA 270
11.6 Combining SSA and SEA in Time Varying MAS Topology 273
11.6.1 The Control Algorithm 274
11.6.2 The Learning Algorithm 275
11.6.3 Performance 275
11.7 Discussions 276
11.7.1 The Energy Based Methods 277
11.7.2 Limitations and Future Work 277
11.8 Conclusion 278
References 278
12 When Human Visual Performance Is Imperfect---How to Optimize the Collaboration Between One Human Operator and Multiple Field Robots 281
12.1 Introduction 281
12.2 Human and Robot Performance in Target Classification [4] 283
12.3 Optimizing Human--Robot Collaboration for Target Classification 285
12.3.1 Predetermined Site Allocation 285
12.3.2 Optimized Site Allocation 290
12.4 Numerical Results 293
12.4.1 Collaboration Between the Human Operator and One Robot [14] 294
12.4.2 Predetermined Site Allocation 296
12.4.3 Optimized Site Allocation 303
12.5 Conclusions 308
References 308
13 Human-Collaborative Schemes in the Motion Control of Single and Multiple Mobile Robots 310
13.1 Introduction 310
13.2 Modeling of the Robot and the Interactions 311
13.2.1 Mobile Robot 311
13.2.2 Communication Infrastructure 313
13.2.3 Human--Robot Interface 313
13.3 A Taxonomy of Collaborative Human--Robot Control 315
13.3.1 Physical Domain of the Robots 315
13.3.2 Degree of Autonomy from the Human Operator 316
13.3.3 Force Interaction with the Operator 319
13.3.4 Near-Operation Versus Teleoperation 321
13.3.5 Physical Interaction with the Environment 322
13.3.6 Use of Onboard Sensors Only 324
13.4 A Taxonomy of Collaborative Human--Multi-robot Control 325
13.4.1 Level of Centralization 325
13.4.2 Master--Leader--Followers Schemes 326
13.4.3 Formation-Orthogonal Control Schemes 327
13.4.4 Group-Property Preservation Schemes 328
13.4.5 Physical Interaction with Contact 329
13.5 Conclusions 330
References 331
14 A Passivity-Based Approach to Human--Swarm Collaboration and Passivity Analysis of Human Operators 334
14.1 Introduction 334
14.2 Intended Scenario and Control Goals 337
14.3 Control Architecture and Passivity 340
14.4 Convergence Analysis 342
14.4.1 Synchronization in Position Control Mode 342
14.4.2 Synchronization in Velocity Control Mode 346
14.5 Passivity of the Human Operator Decision Process 349
14.5.1 Experimental Setup and Approach 350
14.5.2 Analysis on Human Passivity in Position Control Mode 353
14.5.3 Analysis on Human Passivity in Velocity Control Mode 356
14.5.4 Analysis on Individual Variability 358
14.6 Summary 362
References 363
15 Human--Swarm Interactions via Coverage of Time-Varying Densities 365
15.1 Introduction 365
15.2 Human--Swarm Interactions via Coverage 367
15.2.1 The Coverage Problem 370
15.2.2 Centralized Coverage of Time-Varying Densities 373
15.2.3 Distributed Coverage of Time-Varying Densities 377
15.3 Designing Density Functions 382
15.3.1 Diffusion of Drawn Geometric Configurations 382
15.3.2 Control of Gaussian Functions 384
15.4 Robotic Experiments 385
15.5 Conclusions 388
References 389
16 Co-design of Control and Scheduling for Human--Swarm Collaboration Systems Based on Mutual Trust 394
16.1 Introduction 394
16.2 Swarm Setup 397
16.2.1 Dynamic Timing Model and Collaboration Delay 397
16.2.2 Cooperative Control for Swarm Agents 399
16.3 Collaboration Framework 405
16.3.1 Trust Model 405
16.3.2 Human Performance Model 406
16.3.3 Swarm Performance Model 407
16.3.4 Human Attention Preference 407
16.3.5 Fitness 409
16.4 Real-Time Scheduling 410
16.5 Simulation Results 411
16.5.1 Parameter Setup 411
16.5.2 Results and Discussions 413
16.6 Conclusions 417
References 418
Index 421