Dynamics identification and implementation for enhanced haptic display

This report presents a new method for dynamics identification and implements the identified models for enhanced haptic display. The report also derives an adaptive virtual coupling design technique for increasing the control performance in haptic displays. Practical dynamics models are essential for haptic display for Virtual Reality systems. Such dynamic models are desirably acquired via experimental identifications. Having said that, traditional dynamics identification methods usually need large sized training data sets, which maybe hard to meet in several practical applications. To obtain reality based models, we present in this report an identification method using Support Vector Machines (SVM) regression algorithm which is more efficient than traditional methods for sparse training data. This technique hasn’t been formerly studied experimentally in the dynamics identification literature. SVM can be utilized as a generic learning machine or as a special learning technique that can take advantage of the available knowledge about the dynamics structure….


1. Introduction
1.1. Fundamental Issues
1.1.1 Haptic display and Virtual Reality
1.1.2 Haptic Interface Design
1.1.3 Force Control for Haptics
1.2. Dynamics Modeling for Haptic Display
1.2.1 Previous Work in Dynamics Modeling
1.2.2 Friction modeling
1.2.3 Haptic Interface Dynamics Modeling
1.2.4 Some Observations
1.3 Dynamics Identification from Measurement
1.4 Thesis Contributions
1.5 Thesis Outline
2. Support Vector Machines for Dynamics Identification
2.1 Statistical Learning Basis
2.2 Linear Dynamics Identification
2.2.1 Theoretical Basis
2.2.2 Model Identification Method
2.3 Nonlinear Dynamics Identification
2.3.1 Feature Maps by Kernels
2.3.2 Fundamental properties of kernels
2.4. From SVM to Regularization Networks
2.5 ESVM Algorithm for Dynamics Identification
2.6 Algorithm Implementation
2.6.1 Optimal Conditions for SVM
2.6.2 Optimization Methods
2.7 Considerations in SVM implementation
2.7.1 kernel and regularization parameter
2.7.2. The Bias Parameter m
2.8 Summary
3. Contact Friction and Dynamics Model Identification
3.1 Limitation of Existing Contact Friction Models
3.2 Using SVM for Dynamics Identifications
3.3 Friction Identification Strategy
3.3.1 Hard surface friction identification
3.3.2 Soft surface friction identification
3.3.3 Needle Puncture Dynamics Identification
3.4 Experiments
3.4.1 Experimental Set-up
3.4.2 Experimental 1: Hard Surface Friction Identification
3.4.3 Experimental 2: Soft Surface Friction Identification
3.4.4 Experimental 3: Needle Puncture Dynamics Identification
3.5 Summary
4. Interface Device Dynamics Model Identification
4.1 Dynamics characteristics of a mechanical Haptic interface
4.2 Haptic Interface Dynamics Model
4.3 Model Identification
4.3.1 Joint Friction Characteristics
4.3.2 Inertia Identification
4.3.3 Joint Friction Model Identification by ESVM
4.4 Summary
5. Adaptive Virtual Coupling Design for Enhanced Haptic Display
5.1 Network Based Haptic Display with Virtual Coupling
5.2 Adaptive Virtual Coupling Design for Haptic Display
5.3 Stability condition of the haptic display system
5.4 Simulation and Experiments
5.5 Summary
6. Haptic Display Implementation
6.1 Incorporating Dynamics for Enhanced Haptic Display
6.1.1 Two-port Network Based Haptic Controller
6.1.2 Performance Comparison
6.2 Virtual Friction Force Display
6.3 Virtual Needle puncture…

Source: City University of Hong Kong

Leave a Comment