Custom Pluggable State Machine Prototype

This prototype explores the design of a fully customizable pluggable state machine framework built for flexible AI and gameplay behavior management. The architecture allows developers to define states, actions, and decisions independently using Scriptable Objects, enabling reusable and highly modular logic systems.

The framework was designed to separate behavior configuration from implementation code, making it easier to iterate, extend, and debug complex AI systems directly inside the Unity Editor.

The prototype demonstrates how Scriptable Object-driven workflows can improve scalability and reduce tight coupling between gameplay systems while providing a clean visual structure for behavior creation.

The project focused heavily on modular architecture, extensibility, and reusable gameplay logic patterns.

Project overview

This prototype features NPC tanks that navigate a maze environment. The behavior of these tanks is controlled by a state machine which adjusts their actions based on environmental inputs:

• Patrol: Navigate through the maze following a set path.
• Chase: Seek out the player upon detection.
• Attack: Engage the player when within optimal range.

Key features

• Fully custom framework: Using Scriptable Objects, configure and modify behaviors, decisions, and states dynamically without modifying core system code.
• Dynamic AI behaviors: Enable tanks to change behaviors based on gameplay, providing a rich and interactive player experience.

System components

The state machine consists of:

• States: What the AI is currently doing (for example, patrolling, chasing).
• Actions: Tasks the AI performs while in a state (for example, move to a point, look around).
• Decisions: Criteria that trigger transitions between states (for example, player detected, health low).