Currently, l.o.d. is done for the entire terrain to demonstrate the mechanism. In order for l.o.d. to be effective in promoting more effective redering, we must implement l.o.d. for local sections of terrain.

Sometimes the balloon "lands" in midair. This may be because the terrain collision detection method fails due to imprecision of calculations internal to WorldUp when the balloon is over a border between polygons. (The collision detection method uses two rays which are geometrically collinear and should always intersect the same terrain polygon during free flight. Imprecision in the ray/polygon intersection calculation may allow the rays to intersect different polygons in borderline cases. A possible solution is to compare geometries intersected rather than polygons intersected, depending on how geometries are specified.)

Reduction in mass by burning fuel is not modeled. This aspect of simulation is not essential to learning balloon maneuvering or evident in short demo flight sequences, such as this model is most likely to be used for. In a more advance simulator that might be used for teaching a full flight profile, more detailed models for fuel consumption, tank pressure, etc. would be required, and calculated change in mass of fuel would be available as a trivial side-effect of that model.

Currently, varrying wind conditions are generated by manual updates during run time. At best, this means it is essentially impossible for the pilot to be surprised by a change of wind.

Return to main home page http://www-VRL.umich.edu/~saha/eng477/ Last update: October 14, 1996 by DS