Johann Jungbauer (MScEng-I).

Johann Jungbauer
Evolving Squad Behaviour in First Person Shooter AI Agents

Artificial intelligence (AI) is about making computers perform thinking tasks that people and animals do all the time. Examples of these thinking asks would be things like finding your way around (navigation) and making decisions on actions you want to take. In a gaming context AI is more focused on making believable opponents for people to play against. The AI agents in games still need to do thinking tasks it’s just that the digital worlds they inhabit are more simple than our own, e.g. run then shoot an enemy versus what should I make for dinner.

My research is specifically looking at creating AI that can perform group / squad tactics such as flanking and distracting an enemy. Instead of coding the tactics into the AI, I will be using genetic algorithms to generate, try and cull, approaches to posed scenarios. The type of game I will be focussing on will be First-Person-Shooters (FPS). This type of game requires the AI to function in a three dimensional (3D) world, which poses its own challenges. As part of the research, I will be comparing hard-coded AI agents to my grown agents. Eventually, I would like human players to play against my AI agents and provide me with feedback on the experience. This information would help me make judgements on the algorithmic approach as compared to hard-coded AI common in current games.

Read Johann’s blog posts
Join Johann in 2012 at the MIH Media LabĀ 

My contact details: jungbauer@gmail.com