This thesis focuses on virtual solid body manipulation with haptic feedback at real-time frame rates, such as teeth drilling and bone sawing.
Medical operations and surgical interventions require an enormous amount of skill and experience.
The traditional ways for medical students to obtain their surgical skills are either practicing on cadavers, animals or patients directly. However, all of these training methods bear several disadvantages.
With ongoing developments in the area of virtual reality (VR), virtual surgery simulators turned out to be a safe and repeatable alternative to the traditional training methods. A medical simulator provides a virtual training environment in which the trainee is able to interact with virtual objects and obtain physically plausible force feedback via a haptic feedback device connected to the simulator.
Table of Contents
1. Introduction
2. Preliminaries
2.1 Object Representation
2.1.1 Voxels
2.1.2 Marching Cubes Algorithm
2.1.3 Point Shells
2.1.4 Polygonal Meshes
2.1.5 Distance Fields
2.2 Haptic Rendering
2.2.1 Collision Detection
2.2.2 Collision Response
2.2.3 Direct Rendering vs. Virtual Coupling
2.3 Rendering Frequencies
3. Related Work
4. Methods and Implementation
4.1 Object Representation
4.1.1 Dental Tools
4.1.2 Volume Object
4.1.3 Implementation
4.1.3.1 Dental Tools
4.1.3.2 Volume Object
4.2 Collision Detection
4.2.1 Spatial Hashing
4.2.2 Dental Bur
4.2.3 Dental Saw
4.3 Collision Response
4.3.1 Handling Collisions
4.3.2 Erosion
4.3.3 Haptic Device
5. Evaluation
5.1 Voxel Cube
5.2 Tooth and Bur
5.2.1 Setup
5.2.2 Performance
5.2.3 Simulation Quality
5.3 Mandible and Saw
5.3.1 Setup
5.3.2 Performance
5.3.3 Simulation Quality
6. Conclusion and Future Work
6.1 Conclusion
6.2 Future Work
Objectives and Topics
This thesis aims to develop a real-time virtual surgery simulator for dental training, specifically focusing on teeth drilling and mandible sawing operations with haptic feedback. The primary research goal is to implement a system that achieves stable, physically plausible interaction at high haptic update rates (approx. 1000 Hz) while maintaining a separate visual thread for smooth rendering.
- Volumetric object representation using voxel grids and point shells.
- Advanced collision detection techniques, specifically spatial hashing, for real-time performance.
- Collision response using penalty-based methods with viscoelastic spring models.
- Integration of haptic devices with virtual coupling to ensure stability during tool-object interaction.
- Evaluation of performance, simulation quality, and system stability under different scenarios.
Extract from the book
4.2.1 Spatial Hashing
The broad phase step is based on a simplified implementation of a spatial hashing technique. The main idea of this approach is to reduce the amount of point shell points which have to be tested against the signed distance field of the dental tool in use. In order to achieve this, the voxel grid is partitioned into several cells of equal size. Thus, each cell only covers a specific number of voxels and therefore only a part of the complete point shell (Figure 4.5). Now, the idea is to determine all cells the virtual tool in use intersects with and only use the point shell points contained in these cells for further tests.
Every single cell is implemented as a set of references to the point shell points. All together they form a so called hashtable. In order to address a specific cell, they are assigned a unique key and inserted in a map which represents the hashtable. In this case, the key is a three-dimensional vector of integers which determines the position of every cell in the voxel grid. Figure 4.6a shows how the cells would be addressed in two-dimensional space, but this can easily be transferred to three-dimensional space. Hence, every time a new point shell point is calculated, it is put into one of these cells.
Summary of Chapters
1. Introduction: Discusses the necessity of virtual surgery simulators as a safe, repeatable alternative to traditional training and outlines the core requirements for haptic rendering and real-time performance.
2. Preliminaries: Provides the theoretical background on 3D object modeling (voxels, meshes, distance fields) and the principles of haptic rendering, including collision handling and system stability.
3. Related Work: Reviews existing literature and prototype simulators for dental drilling and bone sawing, comparing various voxel-based and surface-based approaches.
4. Methods and Implementation: Details the development of the simulator, covering object representations, the integration of collision detection (spatial hashing), collision response (penalty-based), and the use of virtual coupling.
5. Evaluation: Benchmarks the implementation in two scenarios (tooth drilling and mandible cutting) based on computational performance, simulation frequency stability, and subjective quality assessment.
6. Conclusion and Future Work: Summarizes the thesis results, confirming that the developed simulator fulfills training requirements, and suggests future improvements like enhanced spatial hashing and improved haptic force calculation for saw edges.
Keywords
Virtual Reality, Dental Surgery Simulator, Haptic Rendering, Collision Detection, Spatial Hashing, Voxel-based Representation, Point Shells, Real-time Simulation, Force Feedback, Virtual Coupling, Mandible Cutting, Teeth Drilling, Marching Cubes, Penalty-based Collision, Surgical Training
Frequently Asked Questions
What is the primary purpose of this research?
The research focuses on creating a virtual reality dental surgery simulator that provides a safe and repeatable training environment for medical students, specifically for drilling teeth and cutting mandibles.
What are the core technical challenges addressed?
The main challenge is achieving real-time interaction (approx. 1000 Hz) for haptic feedback while simultaneously rendering high-quality 3D graphics, requiring efficient collision detection and multi-threaded synchronization.
Which scientific methods are employed for simulation?
The study utilizes volumetric modeling via voxel grids and point shells, spatial hashing for accelerated collision detection, and penalty-based collision response using viscoelastic springs and virtual coupling.
What is the role of spatial hashing in this simulator?
Spatial hashing acts as a broad-phase collision detection mechanism that partitions the voxel grid into cells, significantly reducing the number of point shell points that need to be tested against the dental tools.
How is the haptic device connected to the virtual tools?
The simulator implements a virtual coupling approach, where the haptic device and the virtual tool are separated but connected via a viscoelastic spring to ensure system stability and provide realistic force feedback.
What characterizes the performance of the implemented simulator?
The evaluation shows that the simulator successfully maintains a frequency of around 1000 Hz in both the drilling and cutting scenarios, successfully meeting the requirements for stable haptic interaction.
How does the erosion process update the visual model?
The erosion process dynamically decreases the density values of voxels upon collision and updates the surface representation using the Marching Cubes algorithm in the graphics loop.
What limitations were identified for future work?
Future improvements include optimizing spatial hashing for better performance, handling collision normal calculation for the edges of the saw more accurately, and incorporating more complex models or error-tracking systems.
- Citar trabajo
- Thomas Conraths (Autor), 2015, Developing a Dental Surgery Simulator for Teeth Drilling and Mandible Cutting, Múnich, GRIN Verlag, https://www.grin.com/document/1194359
