Collision Detecting Strategies for Virtual Construction Simulation
Date Issued
2008
Date
2008
Author(s)
Lai, Kuan-Chen
Abstract
Collision detection is one of the most important computational methods required to simulate virtual construction on computers. Similarly to an actual construction site, all collisions need to be avoided in the virtual world during the construction process. This means that the computer program which generates the virtual construction scenarios needs to continuously check the collision status of all objects. Although computing the collision status of an object may not consume much time using modern computer equipment, when the construction site is complex or the number of objects on the construction site is large, the computational cost for checking the collisions can be very high. Therefore, this research aims to find efficient methods for performing the collision detection.he computational efficiency of a collision detection method depends highly on the application scenario and precision requirements. Some collision detection algorithms can determine the collision status (i.e. whether or not a collision has occurred) quickly with multiple logic statements while other algorithms can return both the collision status and the shortest distance to the next collision in the case of a collision free event. n this research, we approximate construction machineries and structural elements on a construction site by using spheres and cylinders. By properly modeling the object using primitive shapes, the computational cost for collision detection can be significantly reduced. The VC-COLLIDE algorithm has been developed in this research. It has the following major advantages: (1) it returns not only the collision status but also the distance to a collision; (2) spherical and cylindrical box combinations provides more efficient choices; and (3) it takes note of the construction scenario, and chooses the best collision detection solution for the specific virtual construction scenario. C-COLLIDE has been implemented and tested by using three construction scenarios: small building scenarios (683 objects), large building scenarios (2143 objects) and tall building scenarios (2612 objects). Open Dynamic Engine (ODE), a widely used physics engine, is also involved in the tests. The test results indicate that VC-COLLIDE can complete all the collision checks within 1/20th of a second, the upper bound of real-time refresh time in the three testing environments. The results also indicate that the performance of VC-COLLIDE is significantly better than the performance using ODE. n the future, VC-COLLIDE can be used as a basic computation kernel to render a virtual construction site in real time. It can also support studies on erection simulations, motion planning of construction machines and site layout planning, all of which require efficient collision detection methods to visualize dynamic construction scenarios.
Subjects
construction crane
collision detection
virtual reality
construction simulation
bounding volume box
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
index.html
Size
23.27 KB
Format
HTML
Checksum
(MD5):64f9fa6e641a2eda736332716fe556b5