[theprgramerdude]

Originally Posted by Ether

The paper sets up the differential equations for the flight of a frisbee. The equations are nonlinear; there is no analytical solution ("formula").

Since there is no analytical solution, the paper presents a numerical solution to the initial value problem.

You wouldn't have the necessary computing resources to do this numerical solution in real time on your robot, let alone the considerably more difficult boundary value problem you are asking.

Even if you could, it would be pointless because you would have to validate the model by doing extensive testing to determine the values of all the coefficients. As long as you have to do that, you might as well just use the test data to determine what launch angles and wheel speeds give you the desired flight paths and then just store that info in a lookup table that your code would use to choose the launch parameters.

On what basis do you claim these calculations can't be handled in real-time?