BOTS in FPS Games

BOTS are virtual humans (embodied virtual agents) in FPS (First Person Shooter) games. For instance you can imagine computer opponents in Unreal Tournament, Quake or Half-Life. These bots run, shoot, fight and they need to be controled.

The overall method which is used to control them is called artifical intelligence (AI). As there are many approaches in the field of AI, there are many ways to complete the tasks of bot control. These approaches differs in their usability to solve different types of problems, their universality and also in their computational complexity. Most common approaches are (considering mainly BOTS AI):

  • scripted AI (used in older computer games very often, it can be argued if this is still AI)
  • finite state machines (FSM) (used in computer games almost exclusively now, thanks to its universality)
  • if-then rules (simple approach, close to finite state machines - often FSM are implemented with if-then rules )
  • fuzzy systems
  • planning (used in more in academic sphere)
  • neural nets (can be used for learning, or some well defined problem)

Our group is interested in artificial inteligence, more specifically in creating the control structures of BOTS in FPS games. There are many related problems with this topic. We already finished projects concerning map representation, team AI and artificial emotions. Our big goal is to further pursue AI topics related with FPS bots.

guidelines/bots_in_fps_games.txt · Last modified: 2011/12/22 15:13 by michal.bida