Modular AI is an inspector based visual behaviour designer.
- Implements competing utility theory behaviours for dynamic AI behaviours.
- Zero string comparisons or calls to reflection.
- Behaviours can be fully configured during runtime.
- Not tied to a specific pathfinding solution.
- Fully extendable via code modules.
- Flee/Follow
- Wander
- Patrol
- Remember players, AI agents, or environment objects.
- Focus/target objects in memory
- Invoke custom events.
- Eye-level scans
- Physics Scan Options for 2D & 3D
- Integrated with memory
- Idle, Walk, Run & Jump
- Attack, Death Animations
- Dynamic Hit reactions
All components can be found under Kitbashery in the component menu:
Online documentation & scripting API is found at:
https://kitbashery.com/docs/modular-ai
Modular AI uses utility theory for its AI behaviour logic. An AI agent can have as many behaviours as you want.
Note: Since v.2.0.2 Modular AI supports linear state-machine-like evaluation by default with optional competing utility theory behaviours.
Behaviours are comprised of conditions and actions and have a score value. The behaviour with the score that best meets the score type you set will execute its actions.
Conditions are true/false statements based on what the AI knows about the game world. If a condition meets its desired state then it will add its score to the behaviour's total score.
Actions are executed in the order they are arranged if a behaviour's total score meets the score type better than any other behaviour.
Take a look at the docs for ExampleModule.cs for how to create your own modules.
The name Kitbashery & all associated images Copyright © 2023 Kitbashery. All Rights Reserved.