Ok, so it's not exactly skynet, but I have got my first AI state working, kind of.
The first state is "HOLD" in which case the agent stays in place where they are and shoots at any unit that comes in range. When I started writing this module, I found that the existing method of triggering actions wasn't good enough to allow the AI to choose the best weapon or target. It worked by simply sending a command to the unit to trigger the currently selected action.
If the action is valid, it triggered, if not it didn't.
That's fine for play controlled units, as that's all they need to do. But AI needs to know in advance if the action is valid. The player can get that info from UI feedback, but that wasn't available to the AI player.
There were three problems:
1. The UI feedback duplicated code in the action trigger function. These two sets of code could get out of phase so that UI feedback was wrong.
2. The action trigger didn't give enough feedback for the AI to make decisions.
3. The action trigger method as monolithic. It contained both validation and triggering code.
So I separated the validation part and the triggering part in to two different methods. The UI could now call the validation part to display feedback (check the top left of the animations here), and the AI can use it to get feedback about its combat options too and make choices about which is the best attack to use, or which is the best target.
Now I feel like I'm in a much better position to move forward with programming the AI.
It's not going to be a richly featured generic AI. Instead the mission editor will allow the level designer to give units AI flags to control their behavior.
An "ATTACK" flag will be given to units which are supposed to be assaulting a "DEFEND" objective. They will move towards the target, trying to make use of cover until they get within range of the objective, or they see a player controlled unit. Then they will engage in combat.
Jobs such as support, supply, repair or air support liaison will be handled by units with special flags.
Once I start writing code for random missions, there will have to be a function to assign these flags automatically. But for the first version of the game, the ALPHA build, missions will be written by hand using the fully featured mission editor I designed earlier.
I've added the ability to add objectives, way points and area triggers, as well as another menu for adding AI flags and special conditions, so that units can start the game with their crew knocked out (ready for salvage) or partly damaged (ready for rescue).
Objectives will be color coded, which corresponds to number indexes 1-8. Units can be given an objective index too. Destroy all units associated with "DEFEND" objective to set that objective to "COMPLETE". Escort all vehicles with linked to an "ESCORT" objective to chalk up another victory. And so on.
You can also see an "L" shaped arrow above the red objective above. This is a trigger index. The red objective will only be activated when the Pink objective is "COMPLETE". This allows objectives to be staged as a sequence. A mission might have multiple waves of enemy reinforcements, or it might involve a scouting phase before friendly reinforcements arrive.
Objectives can have a time limit or time delay, and they can be visible or invisible. Some objectives like "NO AIR SUPPORT" are global, though they could be given a trigger index or a time limit to limit their effect.
The first state is "HOLD" in which case the agent stays in place where they are and shoots at any unit that comes in range. When I started writing this module, I found that the existing method of triggering actions wasn't good enough to allow the AI to choose the best weapon or target. It worked by simply sending a command to the unit to trigger the currently selected action.
If the action is valid, it triggered, if not it didn't.
That's fine for play controlled units, as that's all they need to do. But AI needs to know in advance if the action is valid. The player can get that info from UI feedback, but that wasn't available to the AI player.
There were three problems:
1. The UI feedback duplicated code in the action trigger function. These two sets of code could get out of phase so that UI feedback was wrong.
2. The action trigger didn't give enough feedback for the AI to make decisions.
3. The action trigger method as monolithic. It contained both validation and triggering code.
So I separated the validation part and the triggering part in to two different methods. The UI could now call the validation part to display feedback (check the top left of the animations here), and the AI can use it to get feedback about its combat options too and make choices about which is the best attack to use, or which is the best target.
Now I feel like I'm in a much better position to move forward with programming the AI.
It's not going to be a richly featured generic AI. Instead the mission editor will allow the level designer to give units AI flags to control their behavior.
An "ATTACK" flag will be given to units which are supposed to be assaulting a "DEFEND" objective. They will move towards the target, trying to make use of cover until they get within range of the objective, or they see a player controlled unit. Then they will engage in combat.
Jobs such as support, supply, repair or air support liaison will be handled by units with special flags.
Once I start writing code for random missions, there will have to be a function to assign these flags automatically. But for the first version of the game, the ALPHA build, missions will be written by hand using the fully featured mission editor I designed earlier.
I've added the ability to add objectives, way points and area triggers, as well as another menu for adding AI flags and special conditions, so that units can start the game with their crew knocked out (ready for salvage) or partly damaged (ready for rescue).
Objectives will be color coded, which corresponds to number indexes 1-8. Units can be given an objective index too. Destroy all units associated with "DEFEND" objective to set that objective to "COMPLETE". Escort all vehicles with linked to an "ESCORT" objective to chalk up another victory. And so on.
You can also see an "L" shaped arrow above the red objective above. This is a trigger index. The red objective will only be activated when the Pink objective is "COMPLETE". This allows objectives to be staged as a sequence. A mission might have multiple waves of enemy reinforcements, or it might involve a scouting phase before friendly reinforcements arrive.
Objectives can have a time limit or time delay, and they can be visible or invisible. Some objectives like "NO AIR SUPPORT" are global, though they could be given a trigger index or a time limit to limit their effect.
Comments
Post a Comment