Hello everyone,
I was wondering what kind of AI system(s) Splinter Cell Conviction might have used. Just a simple FSM that always has the enemies be in states like "relaxed", "alerted", "seeking player", "combat", "taking cover" at a time? Or did they implement a rather elaborated goal-oriented AI system, that constantly tries to satisfy the enemies's needs?
Greetings,
Shadax
Splinter Cell Conviction AI
I can check with my sources on this, but it is likely NOT a FSM. While there might be over-arching branches like the ones you describe, there are far too many possible actions they can take to reasonably be handled in a straight FSM. I'm quite sure it isn't a planner-based system. It might be goal-oriented to some degree, but I doubt that is the primary component. Figure it is a behavior tree until further notice.
Incidentally, I just finished playing the game the other day. My Post-Play'em blog (below) has two posts on a couple of my observations.
Incidentally, I just finished playing the game the other day. My Post-Play'em blog (below) has two posts on a couple of my observations.
Author
Hello
Thank you for your effort so far, Dave. I assume "the suits" haven't given their approval yet.
Out of curiosity: why do you think that it's probably a BT and not a goal-oriented behavior as the major component? (Disclaimer: I'm just learning about BTs, nevertheless I think I'm used to GOB, at least when it comes to what Mat Buckland's famous book tells us about it).
From my level of knowledge, GOB aims at continuously satisfying all "needs" of the character, thus building a small plan for the most insistent need at a given time, executing the actions set up by the plan until these actions are all successfully finished or until an even more urgent need from the character's state space arises. Thus, I can see many parallels between what I think GOB is about and how the enemies in Splinter Cell behave - at least too many to make my current lack of experience about BTs prove me wrong.
Greetings,
Shadax
Thank you for your effort so far, Dave. I assume "the suits" haven't given their approval yet.
Out of curiosity: why do you think that it's probably a BT and not a goal-oriented behavior as the major component? (Disclaimer: I'm just learning about BTs, nevertheless I think I'm used to GOB, at least when it comes to what Mat Buckland's famous book tells us about it).
From my level of knowledge, GOB aims at continuously satisfying all "needs" of the character, thus building a small plan for the most insistent need at a given time, executing the actions set up by the plan until these actions are all successfully finished or until an even more urgent need from the character's state space arises. Thus, I can see many parallels between what I think GOB is about and how the enemies in Splinter Cell behave - at least too many to make my current lack of experience about BTs prove me wrong.
Greetings,
Shadax
Not sure... but I figure that everything that I was seeing in the game could be covered by a BT. Why get more complicated? I really don't know. I just got back from a vacation so I'll ping my pal and see what's up.
OK... from Lead AI dude, Martin Walsh:
[color="#1F497D"]
[color="#1F497D"]
We actually use a hierarchical FSM (with a couple of twists on how weset up transitions between states that give it a bit of addedfunctionality.) We also have a tactical layer that can assign roles toNPCs based on the game state and those roles also affect state transitions.[/quote]
In retrospect, having recently seen the game, that makes sense. Like I said before, I didn't see a lot of elaborate adaptive behavior that I would expect from a planner.
Author
Thank you for this information, Dave. It should point me in the right direction for learning something about HFSMs by a practical example.
Greetings,
Shadax
Greetings,
Shadax
I'm curious as to why you would think that would be such a help? I assume you plan on trying to figure out what the states and their transitions are. I'm not sure that exercise is going to prove to be anything more than a simple mind game. You really aren't "learning something about HFSMs" at all.
Author
I was hoping that if I get pointed in the right direction I could learn something new about game AI programming. My idea was: first read about the according theory in books and then try to match that theory with a game that used that specific technique successfully. I thought learning something new might be easier that way (I want to prevent reading the "wrong" stuff and getting lost). Does this make sense?
Oh sure, observing things is a great way to polish off the knowledge of what is happening under the hood. I think that you might be thinking that observing states, etc. is a little more complex than it really is, though.
The thing is, a BT, planner, FSM, HFSM, or utility-based architecture all end up in the same place... a "state" or a "behavior" or an "action"... whatever you want to call it. They are doing something at the moment. The only difference between them really is how they get to that point.
A BT is, at its core, a FSM with the transition logic lifted out of the states and put into a manager of some sort in order to make it a lot more organized and manageable. The same can be said for a utility-based system. Now the way that BTs and utility-based systems manage those transitions is significantly different... and that's where the power is. The problem now is that you are starting to get hard pressed to be able to observe what is happening in the decision logic. For the most part, all you can see is that there are situations and states.
Note the sticky topic on the forum about a reading list, by the way.
The thing is, a BT, planner, FSM, HFSM, or utility-based architecture all end up in the same place... a "state" or a "behavior" or an "action"... whatever you want to call it. They are doing something at the moment. The only difference between them really is how they get to that point.
A BT is, at its core, a FSM with the transition logic lifted out of the states and put into a manager of some sort in order to make it a lot more organized and manageable. The same can be said for a utility-based system. Now the way that BTs and utility-based systems manage those transitions is significantly different... and that's where the power is. The problem now is that you are starting to get hard pressed to be able to observe what is happening in the decision logic. For the most part, all you can see is that there are situations and states.
Note the sticky topic on the forum about a reading list, by the way.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
