Robot Learning Human Skills and Intelligent Control Design

Description

In the last decades robots are expected to be of increasing intelligence to deal with a large range of tasks. Especially, robots are supposed to be able to learn manipulation skills from humans. To this end, a number of learning algorithms and techniques have been developed and successfully implemented for various robotic tasks. Among these methods, learning from demonstrations (LfD) enables robots to effectively and efficiently acquire skills by learning from human demonstrators, such that a robot can be quickly programmed to perform a new task. This book introduces recent results on the development of advanced LfD-based learning and control approaches to improve the robot dexterous manipulation. First, there's an introduction to the simulation tools and robot platforms used in the authors' research. In order to enable a robot learning of human-like adaptive skills, the book explains how to transfer a human user’s arm variable stiffness to the robot, based on the online estimation from the muscle electromyography (EMG). Next, the motion and impedance profiles can be both modelled by dynamical movement primitives such that both of them can be planned and generalized for new tasks. Furthermore, the book introduces how to learn the correlation between signals collected from demonstration, i.e., motion trajectory, stiffness profile estimated from EMG and interaction force, using statistical models such as hidden semi-Markov model and Gaussian Mixture Regression. Several widely used human-robot interaction interfaces (such as motion capture-based teleoperation) are presented, which allow a human user to interact with a robot and transfer movements to it in both simulation and real-word environments. Finally, improved performance of robot manipulation resulted from neural network enhanced control strategies is presented. A large number of examples of simulation and experiments of daily life tasks are included in this book to facilitate better understanding of the readers. Cover Half Title Title Page Copyright Page Contents Preface Author Biography Acknowledgements Chapter 1: Introduction 1.1. Overview of sEMG-based stiffness transfer 1.2. Overview of robot learning motion skills from humans 1.3. Overview of robot intelligent control design References Chapter 2: Robot Platforms and Software Systems 2.1. Baxter robot 2.2. Nao robot 2.3. KUKA LBR iiwa robot 2.4. Kinect camera 2.5. MYO Armband 2.6. Leap Motion 2.7. Oculus Rift DK 2 2.8. MATLAB Robotics Toolbox 2.9. CoppeliaSim 2.10. Gazebo References Chapter 3: Human-Robot Stiffness Transfer-Based on sEMG Signals 3.1. Introduction 3.2. Brief introduction of sEMG signals 3.3. Calculation of human arm Jacobian matrix 3.4. Stiffness estimation 3.4.1. Incremental stiffness estimation method 3.4.2. Stochastic perturbation method 3.5. Interface design for stiffness transfer 3.6. Human-robot stiffness mapping 3.7. Stiffness transfer for various tasks 3.7.1. Comparative tests for lifting tasks 3.7.2. Writing tasks 3.7.3. Human-robot-human writing skill transfer 3.7.4. Plugging-in task 3.8. Conclusion References Chapter 4: Learning and Generalization of Variable Impedance Skills 4.1. Introduction 4.2. Overview of the framework 4.3. Trajectory segmentation 4.3.1. Data segmentation using difference method 4.3.2. Beta process autoregressive hidden Markov model 4.4. Trajectory alignment methods 4.5. Dynamical movement primitives 4.6. Modeling of impedance skills 4.7. Experimental study 4.7.1. Learning writing tasks 4.7.2. Pushing tasks 4.7.3. Cutting and lift-place tasks 4.7.4. Water-lifting tasks 4.8. Conclusion References Chapter 5: Learning Human Skills from Multimodal Demonstration 5.1. Introduction 5.2. System Description 5.3. HSMM-GMR Model Description 5.3.1. Data Modeling with HSMM 5.3.2. Task Reproduction with GMR 5.4. Impedance Controller in Task Space 5.5. Experimental Study 5.5.1. Button-pressing Task 5.5.2. Box-pushing Task 5.5.3. Pushing Task 5.5.4. Experimental Analysis 5.6. Conclusion References Chapter 6: Skill Modeling Based on Extreme Learning Machine 6.1. Introduction 6.2. System of teleoperation-based robotic learning 6.2.1. Overview of teleoperation demonstration system 6.2.2. Motion Capture Approach based on Kinect 6.2.3. Measurement of angular velocity by MYO armband 6.2.4. Communication between Kinect and V-REP 6.3. Human/robot joint angle calculation using Kinect camera 6.4. Processing of demonstration data 6.4.1. Dynamic time warping 6.4.2. Kalman Filter 6.4.3. Dragon naturally speaking system for verbal command 6.5. Skill modeling using extreme learning machine 6.6. Experimental study 6.6.1. Motion Capture for Tracking of Human Arm Pose 6.6.2. Teleoperation-Based demonstration in VREP 6.6.3. VR-based teleoperation for task demonstration 6.6.4. Writing Task 6.7. Conclusion References Chapter 7: Neural Network-Enhanced Robot Manipulator Control 7.1. Introduction 7.2. Problem description 7.3. Learning from multiple demonstrations 7.3.1. Gaussian mixture model 7.3.2. Fuzzy Gaussian mixture model 7.4. Neural networks techniques 7.4.1. Radial basis function neural network 7.4.2. Cerebellar model articulation neural networks 7.5. Robot manipulator controller design 7.5.1. NN-based controller for robotic manipulator 7.5.2. Adaptive admittance controller 7.6. Experimental study 7.6.1. Test of the adaptive admittance controller 7.6.2. Test of the NN-based controller 7.6.3. Pouring task 7.7. Conclusion References Index