Techniques for Skeletal-Based Animation in Massive Crowd Simulations

Abstract

Crowd systems play an important role in virtual environment applications, such as those used in entertainment, education, training, and different simulation systems. Performance and scalability are key factors, and it is desirable for crowds to be simulated with as few resources as possible while providing variety and realism for agents. This paper focuses on improving the performance, variety, and usability of crowd animation systems. Performing the blending operation on the Graphics Processing Unit (GPU) side requires no additional memory other than the source and target animation streams and greatly increases the number of agents that can simultaneously transition from one state to another. A time dilation offset feature helps applications with a large number of animation assets and/or agents to achieve sufficient visual quality, variety, and good performance at the same time by moving animation streams between the running and paused states at runtime. Splitting agents into parts not only reduces asset creation costs by eliminating the need to create permutations of skeletons and assets but also allows users to attach parts dynamically to agents.