|Reinforcement Learning Based Character Locomotion in Hitman: Absolution
|IO Interactive / Square Enix
|Track / Format:
Did you know free users get access to 30% of content from the last 2 years?
|Character locomotion systems usually consist of manually crafted blend-trees and large chunks of gameplay code in order to drive and micro-manage them. In this session, we'll talk about the novel approach that we took to create the crowd characters in Hitman: Absolution. This talk will cover how we arranged animation clips in an appropriate data-structure, the controller strategies we used to derive the animation clip selection from high-level AI goals, and how the system can be used in path following scenarios. We will touch on the challenges that we faced during the development of the system.