AI Tutorial Notes

Introduction

• Creating a simple Artificial Intelligence (AI) in Unity.
• AI characters will move in an enclosed area without colliding with each other.
• Many methods exist to achieve this; the chosen method is preferred for its modifiability.

Project Setup

• Software Used: Unity
• New Project: Name it "AI Crow"
  • Select 3D and no packages.
  • Create an empty project.

Initial Configurations

• Unity Settings:
  • Turn off shadows to reduce lag during recording.

Creating the Level

• Plane Creation:
  • Create a plane for AI movement.
  • Expand the plane as needed.
• Folders Setup:
  • Create folders for scripts, scenes, and artwork (for materials).
• Scene Saving:
  • Save the scene as "test".

Creating Boundaries

• 3D Objects:
  • Create boundary cubes to prevent AI from falling off.
  • Duplicate and rotate cubes as needed to form an enclosed area.

Camera Setup

• Move and rotate the camera to get an aerial view of the level.
• Position the camera appropriately on the Z-axis.

AI Character Setup

• Material Creation:
  • Create a material with a darker color for the plane.
  • Create another material for the AI with a red color.
• AI Creation:
  • Create a cube as the AI object, resize it appropriately.
• Attach AI Controller:
  • Create a script named "AIController" and open it in Visual Studio.

Programming the AI Controller

Sensor Creation

• Sensors:
  • Create four sensors for the AI (front, left, right, back).
  • Use Gizmos for visual representation in the editor.

Movement Logic

• AI Movement:
  • Add movement to the AI using transform.position += transform.forward * speed * Time.deltaTime.
  • Implement public float for speed (default value of 10).

Sensor Detection

• Raycast Implementation:
  • Use Raycast to detect obstacles.
  • Implement logic to determine collision with walls.
• Sensors should trigger different movements based on detected obstacles:
  • Front sensor: if wall detected, move backward.
  • Back sensor: if wall detected, move forward.
  • Left and right sensors: adjust turning based on wall detection.

Additional Features

• Turn Value: Control AI direction with turn value.
• Flag Check: Use a flag to detect if a sensor has detected anything.
  • Adjust turn value based on sensor detection.

Final Testing

• Duplicate AI instances to test movement and collision detection.
• Observe AI behavior in response to obstacles.

Summary

• The AI can navigate without colliding by using sensors and simple turning logic.
• Potential for enhancements:
  • Pathfinding, random target selection, or using Unity's NavMesh.
• Project Files: Will be available in the description after the video upload.

Conclusion

• Feedback encouraged for future AI ideas.
• Thanks for following along!