/Training-ML-Agents-

A reinforcement learning project using Unity ML Agents for training interactive sidekick NPC for players

Primary LanguageC#

Training-ML-Agents-

Training a interactive friendly buddy NPC for players in games using Unity ML-Agents reinforcement learning.

Demonstration of a the current progress (Using a character with softbody physics as AI agent) - https://twitter.com/ABadguj/status/1540704208696971266?s=20&t=oR0qN5UFfi2s7yVymlTrsg

Features added -

  1. Training a agent to follow the player, and stay in a fixed radius from him

Future Goals -

  1. Protect player from enemie's melee and projectiles (currently working on it).
  2. Add various behaviors to it like Idle, Agressive, Defensive etc, training the agent with different parameters and environment to achieve various behaviors. Many more to be updated..