Human-Aware Robotics: Modeling Human Motor Skills for the Design, Planning and Control of a New Generation of Robotic Devices

Foreword
Acknowledgements
Contents
1 Introduction
1.1 How Do We Move? A Very Brief Historical Overview
1.2 Human Motor Control
1.2.1 Is Motion a Reflex or an Active Process?
1.3 The Problem of Dimensionality Reduction
1.4 The Implication for Robotics and the Advancement of Technologies
1.5 Open Questions and Proposed Solutions
1.6 Contents of the Monograph
1.6.1 Novelties
References
Part I Taming the Complexity of Human Motion Generation
2 Understanding the Principal Modes of Natural Movements in Temporal Domain
2.1 Introduction
2.2 Experimental Protocol and Setup
2.2.1 A Set of Daily Living Tasks
2.2.2 An Experimental Setup for Data Acquisition
2.3 Motion Identification
2.3.1 Modeling of Upper Limb Kinematics
2.3.2 Model Parameters
2.3.3 Markers Modeling
2.3.4 Model Calibration and Angles Estimation
2.3.5 Experimental Results
2.4 Data Analysis
2.4.1 Segmentation
2.4.2 Time Warping
2.4.3 Principal Component Analysis
2.4.4 Functional Principal Component Analysis
2.4.5 Movement Reconstruction and Performance Analysis
2.5 Conclusions and Implications for Robotics and Bioengineering
References
3 Quantifying the Time-Invariance Properties of Upper Limb Synergies
3.1 Introduction
3.2 Related Work
3.3 Experimental Setup
3.3.1 Setup and Experiments
3.4 Data Analysis
3.4.1 Dynamic Time Warping
3.4.2 Repeated Principal Component Analysis
3.5 Results
3.5.1 Robustness Across Subjects and Validation
3.5.2 Principal Component Description
3.6 Discussions
3.7 Conclusions
References
4 Evidences on the Hierarchical Control of Human Hands
4.1 Kinematic Domain
4.1.1 Introduction
4.1.2 Materials and Methods
4.1.3 Results
4.1.4 Discussion and Conclusions
4.2 Force Synergies in Environmental Constraint Exploitation
4.2.1 Materials and Methods
4.2.2 Results
4.2.3 Discussions and Conclusions
References
Part II On the Design of Nature-Inspired Prostheses and the Assessment of Motion Impairment
5 Using Nature-Inspired Principles to Design of Robotic Limbs: The Soft Wrist
5.1 Introduction
5.2 A Wrist with Synergies
5.3 Mechanical Design
5.4 Wrist Synergies in Humans
5.4.1 Experimental Procedure
5.4.2 Data Analysis
5.5 Experiments
5.5.1 System Demonstration
5.5.2 Implementation of PCs
5.6 Conclusions
References
6 A Novel Approach to Quantify Motion Impairment
6.1 Introduction
6.2 Experimental Protocol and Setup
6.2.1 Set of Daily Living Tasks
6.2.2 Experimental Setup for Data Acquisition
6.2.3 Study Information
6.3 Data Analysis
6.3.1 Modeling and Pre-processing
6.3.2 Evaluation-Index of Motion Complexity
6.4 Results and Discussions
6.5 Implications and Conclusions
References
7 A Novel Mechatronic System for Evaluating Elbow Muscular Spasticity Relying on Tonic Stretch Reflex Threshold Estimation
7.1 Introduction and Motivation
7.2 Spasticity and Equilibrium Point Hypothesis
7.3 Mechanical Design
7.3.1 Evaluation of Mechanical Strength
7.4 Control
7.4.1 Mode A: Position Control
7.4.2 Mode B: Torque Control
7.5 Preliminary Experiments with Healthy Subject
7.6 Conclusions
References
Part III Transferring Human Principles to Cobots and Autonomous Robots
8 Natural Motion: Embedding Human-Likeliness in Robot Movements
8.1 A New Method to Generate Human-Like …
8.1.1 Introduction
8.1.2 Functional Principal Components of Upper Limb Motion
8.1.3 Proposed Strategy
8.1.4 Simulations
8.1.5 Discussion and Conclusions
8.2 A Control-Based Approach to Motion Mapping
8.2.1 Introduction
8.2.2 Mapping Between Kinematics via Impedance Control
8.2.3 Experiments
8.2.4 Discussions and Conclusions
References
9 A Focus on Motion Dynamics: Planning Impedance Behaviors in Physical Interaction
9.1 Introduction
9.2 Background
9.2.1 Redundant Robots
9.3 Impedance Optimization
9.3.1 Interaction Modeling
9.3.2 Robot Definition
9.3.3 Cost Function
9.3.4 Problem Complexity and Generalization to Parallel Robots
9.4 Results
9.4.1 Extension to Parallel Robots
9.5 Generalization to 3D Robots
9.6 Discussions and Conclusions
References
10 Learning from Humans How to Grasp: A Reactive-Based Approach
10.1 Introduction
10.2 The Method
10.2.1 The Human-Robot Interface and Motion Capture System
10.3 Grasp Primitives
10.3.1 Data Acquisition
10.3.2 Primitive Identification
10.4 Experiments
10.4.1 Implementation
10.4.2 Experiments with Robotic Arm: Handover Task
10.4.3 Experiments with Robotic Hand: Grasping an Object from a Table
10.5 Discussions and Conclusions
References
11 Learning from Humans How to Grasp: Enhancing the Reaching Strategy
11.1 Introduction
11.2 Proposed Approach
11.3 Deep Classifier
11.3.1 Object Detection
11.3.2 Primitive Classification
11.4 Robotic Grasping Primitives
11.4.1 Experimental Setup
11.4.2 Approach Phase
11.4.3 Grasp Phase
11.4.4 Control
11.5 Experimental Results
11.6 Discussion
11.7 Conclusions
References
12 Learning to Prevent Grasp Failure with Soft Hands: From On-Line Prediction to Dual-Arm Grasp Recovery
12.1 Introduction
12.2 Methods
12.3 Results
12.3.1 Validation of the Neural Architecture
12.3.2 Validation of the On-Line Integrated Framework
12.4 Discussions and Conclusions
References
13 Dexterity Augmentation of Robotic Hands: A Study on the Kinetic Domain
13.1 Introduction
13.2 Background
13.2.1 Modeling a Compliant Hand with Synergies
13.2.2 Optimization of Grasping Force Distribution
13.3 Materials and Methods
13.4 Results and Discussion
13.4.1 Synergy Incrementality and Optimal Hand Configuration
13.4.2 The Role of Synergy Hierarchy
13.5 Conclusions
References
14 Exploiting Principal Components for Robots Walking: An Approach for Sub-Optimal Locomotion
14.1 Introduction
14.2 Problem Definition
14.3 Optimal Gait Encoding via Principal Components
14.3.1 Principal Components Analysis
14.3.2 Component Maps
14.4 Experimental Validation
14.4.1 Walking with PCs—Constant Speed
14.4.2 Walking with PCs and with PCs and CMs—Speed Variation
14.4.3 Robot Performance Evaluation
14.5 Conclusions
References
Appendix Conclusions and Lessons Learned