Roller Coaster Simulation
Demo
Usage:
This program is develeoped for Linux environment.
To compile, in terminal and in project folder, type:
make
To clean up the executable, type:
make clean
To run, type:
./RollerCoaster track.txt
Details about Functionalities
- The track are Catmull-Rom splines connected together. I used the recursive subdivision algorithm given in the hint slides to draw splines. One thing to note is that I'm only using the subdivision method to store u's. Then using these as drawing points multiple times for later rendering. The order of the points are from start point to end point because my subdivide() method first recursively call the left half, then the right half. That's the reason that I am abling to move the camera directly from the beginning to end through the stored vector of drawing points. Note that my subdivide does two things: 1) Calculate location for each drawing point and store them. 2) Calculate tangent vector of each drawing point and store them.
- I implemented Sloan's method mentioned in to determine normalized normals of the coaster at every drawing point. These normalized normals are set to be the camera's up vector. This method gives more realistic feel of actually riding on a roller coaster compare to simply set the camera's up vector to be the normalized normal of the ground.
- Rendering of the double rail: The way I achived this is, again, thanks to Sloan's methd. Given tangent vectors and their normals, we can calculate their binormal vectors. Then, for each point on spline, we get a local coordinate system defined by these three unit length vectors. Thus, we can then use -N[i] - B[i] as the direction of moving the cross section to left bottom corner for drawing point i, and -N[i] + B[i] as the direction of moving the cross section to right bottom corner. Then by rendering these two corss sections, we get the double rail. The camera position never changed in process of rendering rails. It always moves along the original generated drawing points.