GPU based crowd simulation and visualization

OData support
Dr. Tóth Balázs György
Department of Control Engineering and Information Technology

The number of calculations required to compute the interactions between entities grows rapidly as the number of entities is increased in a simulation of a crowd of autonomous entities, thus real-time simulation becomes impossible with populations above the several hundreds in systems built on the conventional serial execution model, similarly, the visualization of the results quickly reaches its limits.

It is practical to use the graphics processing unit (GPU) for this problem of parallel nature, furthermore all the necessary data for visualization becomes available on the hardware specialized in rendering without transfer.

In the thesis I present a classical model of animal group behavior, and implement its algorithm executed on, and rendered by the graphics hardware.

Through the steps of building the framework, I demonstrate the effect of modifications aimed at accelerating and extending features on performance in the two-, and later three-dimensional simulation, focusing on the upper limit of the population that allows real-time simulation.


Please sign in to download the files of this thesis.