Tiered Behaviour Architecture for Virtual Characters using Cyclic Scheduling and Behaviour Capture

  • Author / Creator
    Zhao, Richard
  • A story-based video game contains many characters. A few are controlled by the player and the majority are virtual characters, controlled by artificial intelligence. In recent years, video game artificial intelligence has developed slower than other aspects of video games, such as graphics, mainly due to the cost of scripting complex and believable virtual characters. To tackle this bottleneck in content creation, several behaviour architectures have been proposed over the past decades. This dissertation proposes metrics for evaluating behaviour architectures and their associated toolsets: expressiveness, performance, quality, and usability. A good mechanism for evaluating architectures / toolsets is essential to replace the current trial and error approach that is in practice today. This dissertation also proposes a new Tiered Behaviour Architecture model and its associated toolset for controlling the behaviours of virtual characters, and evaluates it by applying the metrics. The objective (top) level of the architecture determines the general schedules of the virtual characters composed of objectives, and the roles that will accomplish these objectives. Cyclic scheduling is a technique that allows for the automatic generation of schedules based on partially specified constraints. The role (bottom) level of the architecture uses Behaviour Capture with Hidden Markov Models, a self-contained technique, to generate actual fine-grained behaviours. This dissertation presents studies that show behaviours generated by this architecture / toolset have high scores for all four metrics. It also shows that Behaviour Capture with Hidden Markov Models generates higher quality (more believable) behaviours for virtual characters than the state of the art in commercial games.

  • Subjects / Keywords
  • Graduation date
    Spring 2015
  • Type of Item
  • Degree
    Doctor of Philosophy
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.
  • Language
  • Institution
    University of Alberta
  • Degree level
  • Department
  • Supervisor / co-supervisor and their department(s)
  • Examining committee members and their departments
    • Mueller, Martin (Computing Science)
    • Bowling, Michael (Computing Science)
    • Riedl, Mark (School of Interactive Computing, Georgia Institute of Technology)
    • Gouglas, Sean (History and Classics)
    • Szafron, Duane (Computing Science)