The floating point origin and AI

Hello everyone.

I felt I should ask this question before i get too bogged down in it and find that the path leads to a dead end.

I’m working on a space simulator and have managed to procedurally create the milkyway, this is pretty useless though atm as I can physically use it for the mpment.

But for creating solar systems, a scale issue came up and i thought a clever workaround would be using multiple cameras and a ratio speed script to recreate the solar system at a 1AU scale and then have the objwcts scale up in the main view as the subscene camera moved near to a planet.

However, this subcamera is slaved to the main camera and the main camera moves at a 1:1 scale which means when moving forward it can cover large distances.

However this still wasn’t a viable solution because of the floating point problems at distance. Searching the net i found a solution called “floating point origin”.

Now i understand that basically this moves everything to you and the camera used stays at 0,0,0. However i have a feeling that this solution would cause problems with AI characters, yet there are no mentions on the net if this is a problem or not.

So my question is, does the floating point origin affect how AI behave?


The answer to this question depends on how you design the scales of your world, and the system that you choose for your AI. This video could help you to solve the first requirement: