State machines for AI behavior in Unity - Step-by-Step Execution

Concept Flow - State machines for AI behavior

Start AI

↓

Check Current State

↓

Execute State Behavior

↓

Check Transition Conditions

↓

Change State

↓

Loop Back to Check Current State

The AI starts, checks its current state, runs that state's behavior, then checks if it should switch states based on conditions, looping continuously.

Execution Sample

Unity

enum State { Idle, Patrol, Chase }
State currentState = State.Idle;

void Update() {
  switch(currentState) {
    case State.Idle: /* wait */ break;
    case State.Patrol: /* move around */ break;
    case State.Chase: /* follow player */ break;
  }
  // Check transitions
}

This code shows a simple AI with three states and how it runs behavior based on the current state each frame.

Execution Table

Step	Current State	Condition Checked	Condition Result	Action Taken	Next State
1	Idle	Player detected?	No	Idle behavior (wait)	Idle
2	Idle	Player detected?	Yes	Switch to Chase	Chase
3	Chase	Player lost?	No	Chase behavior (follow player)	Chase
4	Chase	Player lost?	Yes	Switch to Patrol	Patrol
5	Patrol	Player detected?	No	Patrol behavior (move around)	Patrol
6	Patrol	Player detected?	Yes	Switch to Chase	Chase
7	Chase	Player lost?	No	Chase behavior (follow player)	Chase
Exit	Chase	Game ends or AI disabled	-	Stop AI	-

💡 AI stops when game ends or AI is disabled

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6	After Step 7	Final
currentState	Idle	Idle	Chase	Chase	Patrol	Patrol	Chase	Chase	-

Key Moments - 3 Insights

Why does the AI stay in the same state sometimes instead of switching?

Because the condition to switch states is false (see steps 1, 3, and 5 in the execution_table where the condition result is No, so the AI continues current behavior).

How does the AI decide which state to switch to?

It checks specific conditions for each state (like 'Player detected?' or 'Player lost?') and switches only if those conditions are true, as shown in steps 2, 4, and 6.

What happens if no conditions are met to change state?

The AI keeps executing the behavior of the current state, looping back to check conditions again next update (see steps with 'No' condition results).

Visual Quiz - 3 Questions

Look at the execution_table at step 2. What is the AI's current state after this step?

A Chase

B Idle

C Patrol

D Stopped

💡 Hint

Check the 'Next State' column for step 2 in the execution_table.

At which step does the AI switch from Chase to Patrol?

A Step 3

B Step 5

C Step 4

D Step 6

💡 Hint

Look for the row where 'Current State' is Chase and 'Next State' is Patrol.

If the condition 'Player detected?' is always No, what state will the AI stay in according to variable_tracker?

A Chase

B Idle

C Patrol

D Stops immediately

💡 Hint

See variable_tracker where currentState stays Idle when 'Player detected?' is No.

Concept Snapshot

State machines control AI by defining states and transitions.
Each update, AI runs current state behavior.
Then it checks conditions to switch states.
If condition true, AI changes state; else stays.
This loop repeats, making AI react simply and clearly.

Full Transcript

This visual execution shows how a simple AI uses a state machine to decide what to do. The AI starts in an Idle state, waits, and checks if the player is detected. If yes, it switches to Chase state to follow the player. If the player is lost, it switches to Patrol state to move around. Each step shows the current state, the condition checked, the result, and the next state. Variables track the current state over time. Key moments clarify why the AI sometimes stays in the same state and how it decides to switch. The quiz tests understanding of state changes and conditions. This helps beginners see how state machines make AI behavior organized and easy to follow.