Skip to content

Simple combat AI for Skyrim using computer vision.

Notifications You must be signed in to change notification settings

chktrp/bot-skyrim

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Bot plays Skyrim (simple combat)

This is a bot that can perform a simple combat in Skyrim. It uses SSD lite mobilenet v2 to obtain the enemy locations and acts according to the predefined rules.

Bot description

This is my attempt to create a bot that can perform a simple task in a videogame. Its behavior is simple.

  • It'll move backward if the nearest enemy stays too close.
  • It'll try to center the crosshair to the nearest enemy.
  • It'll try to circle around the enemy
  • It'll attack (hold attack key) the enemy if it stays close enough to the crosshair.
  • It'll turn around for a moment if no enemy is visible.
  • It'll reload a quick-save (press F9 key) if it couldn't find any enemy.

Prerequisites

  1. Skyrim should be run in window-mode with 640x400 resolution, locating in the top-left corner of the screen.

  2. Attack key should be set to X instead of default Left-click

  3. Player character should have Flames equipped on their right hand.

  4. A quick-save at the location where the enemies are around.

  5. Anaconda (or miniconda) as package manager.

Instructions

  1. Install dependencies: conda install --file requirements.txt

  2. Run the bot: python play.py

  3. Swap to Skyrim window after the countdown started (5 seconds). It'll take control the key A, S, D, W, X, F9 and mouse cursor. Be prepared!

  4. It should run for about 30 seconds or until no enemy is detected and terminate itself.

Video

https://www.youtube.com/watch?v=uNbRt3y6bME

Reference

This project has a huge inspiration from https://github.com/Sentdex/pygta5

About

Simple combat AI for Skyrim using computer vision.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages