Learning by experience - autonomous virtual character behavioural animation

  • Authors:
  • Tao Ruan Wan;Wen Tang

  • Affiliations:
  • Department of Electronic Imaging and Media Communications, School of Informatics, University of Bradford, Bradford BD7 1DP, UK;School of Computing and Mathematics, University of Teesside, Middlesbrough, TS1 3BA, UK

  • Venue:
  • Intelligent agents for mobile and virtual media
  • Year:
  • 2002

Quantified Score

Hi-index 0.01

Visualization

Abstract

In this chapter, we present a novel goal-orientated approach for complex virtual character behavioural simulation. Our approach is based on the concept of an artificial brain that is a simulation of real human brain activities for simulating the virtual character's behaviour in a virtual world populated with other virtual objects and characters. The control unit in the simulation system can collect and store all the information that is obtained through the virtual character's complex experience, such as learning how to walk and jump, and analyses the information to find a better solution for a specific task. Therefore the virtual character's skill in a particular task will be developed or evolved. The core techniques also include a physics-based human model for motion modelling, which is driven by muscle forces. This approach therefore produces a more accurate simulation of the real world than conventional methods. We demonstrate this by presenting an implementation of this approach.