Videos

Infinite Mario AI – Long Level



Robin Baumgarten

This is my attempt at a Mario AI using a path-finding algorithm called A*. The bot won both Mario AI competitions this year!

You can see the path it plans to go as a red line, which updates when it detects new obstacles at the right screen border. It uses only information visible on screen.

At the “close call” situation: In this version of Mario, when you’re jumping while sliding on a wall, you jump backwards and upwards away from it. That’s what the AI did twice to get out of the hole. This is possible in this version of Mario, which is a Freeware Java clone. See the links below for more info.

The source code is now available on my homepage!

More info:
Competition: http://julian.togelius.com/mariocompetition2009/index.php
My Project Page: http://aipanic.com/h3h/dokuwiki/doku.php?id=projects:marioai
A star: http://en.wikipedia.org/wiki/A*_search_algorithm
Twitter: https://twitter.com/Robin_B
Enjoy 🙂

*Update*: My Mario AI won the ICE-GIC conference competition this year! Yay!

Source

Similar Posts

42 thoughts on “Infinite Mario AI – Long Level
  1. What are you thoughts on using a Branch and Bound search instead? If I recall A* is only a little bit more efficient and probably takes up more memory

  2. The AI is fuelled by its desire to break the ribbon at the end of the course. It was programmed to understand that breaking the ribbon would launch the nukes.
    Nobody tell it.

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com