AI Adventures #4 - AI overhaul in "Forgotten Memories" - By Michael Marquez
I ran in to quite a few issues with the current iteration of the AI. Let's tear it down and reuse what we can and rewrite what I have in mind.
So, I restarted with a mission in mind and with the knowledge gained through my experimentation with the AI and tokens. Getting the data can be accomplished in multiple ways, I ran with tokens and multicasting. Lots of research and time went into reworking the current version and will be used from here on with some modifications.
Getting the data from data-tables, Task services, AI controller, AI actor, Animation, etc. The list goes on for a while. I had to organize the data in a very generic way with all common classes accounted for.
1. Health - unique to each actor but derived from base class
2. Actor - can be stored in array and exposed for editing
3. Animations/Montages - can be stored in array and exposed for editing
4. Skeletal mesh - can be stored in array and exposed for editing.
5. Location - this is tracked by the engine and can be looked up per actor.
6. Actions - unique to each actor but derived from base class.
7. Patrolling behavior - spline or random or a mix of both
7. Behavior tree - everything is called from the task tree! well the animations are at least, its complicated.
there are more things, but these are the basics. I now have a far more modular version that can take in all the aforementioned items and create a variant based on the base class. It's still a work in progress but getting better and better as we learn more with continued development.
Comments
Post a Comment