Skip to content

This is a ML Unity project that works with ML agents RL to train a flying-trigger enemy NPC.

License

Notifications You must be signed in to change notification settings

Jumalita/Unity_MLAgents-ShootAndDodge

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MLAgents ShootAndDodge-IA

Machine learning project within the Unity environment where we're actively using ML agents reinforcement learning (RL) techniques to develop an enemy NPC character. This character, designed to be an flying-trigger enemy NPC that we're meticulously training to enhance its decision-making skills and adaptability through continuous interactions with its virtual surroundings. Leveraging the cutting-edge capabilities of Unity's ML agents RL framework, we're crafting a complex and dynamic training regimen that facilitates the NPC's ability to learn from its experiences, evolving its strategies for optimal performance.

Feel free to navigate through the repository's different branches to explore the various training phases that showcase the progression of our project.

Characteristics of the Enemy NPC

  • Flight Capability: The enemy character possesses the ability to navigate through the virtual environment using a flying mechanism.
  • Projectile Attack: Armed with offensive capabilities, the enemy NPC engages in combat by launching projectiles at its adversaries.
  • Vulnerability: The enemy NPC is susceptible to defeat upon contact.
  • Comparable Speed: Lower velocity, preventing it from overtaking the player unfairly.
  • Strategic Cooldown: To maintain a fair and engaging gameplay experience, the enemy NPC displays tactical acumen by imposing a cooldown period following each projectile assault, requiring prudent timing for optimal effectiveness.

Phase 1

Learning whether to shoot.

Multimedia1.mp4

Phase 2

Learning with movement.

Multimedia2.mp4

Phase 3

Learning with movement at the NPC and the Player.

Multimedia3.mp4

Phase 4

Learning with movement at the NPC and the Player, now an IA.

Multimedia4.mp4

Phase 5

Learning with movement at the NPC and the Player with better stats.

upcoming soon...

This repository is a Computer Engineering ML project for the subject of Advanced Tecniques of Machine Learning at the University of Girona

Thank you for reading,

Juma 💛

About

This is a ML Unity project that works with ML agents RL to train a flying-trigger enemy NPC.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages