Dissertation project
During my third year of university I successfully completed a dissertation project, gaining an overall mark of 72%. My dissertation project, entitled "Real time crowd simulations incorporating individual agent personality models and group interactions", involved the design and development of a crowd simulation model, I then attempted to incorporate a per-agent personality model, based on the OCEAN personality model directly into the local area obstacle avoidance and locomotion routines of each individual agent.
At the time of publishing, this project marked a new, pioneering attempt to integrate individual agent personality into crowds. While personality models have been built upon existing crowd simulations, and attempts have been made to integrate them at a higher level, at the time this work was produced I was unable to locate any evidence to suggest academic work regarding the implementation of a personality model directly into the navigation and locomotion models of the individual agents within a crowd simulation.
Abstract
This paper presents an approach to per-agent based personality and intelligence into a crowd simulation, in order to create a more varied crowd. The paper first meets the basic requirements for a crowd simulation, focusing on medium density crowds, path planning, locomotion and local area obstacle avoidance algorithms are first implemented. The paper then moves on to investigate and implement ways in which agents within the simulation can be given personality and intelligence. The experiments performed on the simulation proved that it is possible to vary agents using variables mapped to OCEAN variables to create a more dynamic simulation.
Keywords
Crowd Simulation, autonomous agents, human behaviour models, artificial personality, OCEAN
Latest blog posts:
