An Intelligent Action Algorithm for Virtual Human Agents

ABSTRACT: The objective of this study is to develop an intelligent action algorithm for virtual human agents on three-dimensional large terrains to accomplish a specified mission by group communication and coordination. The area contains natural and build-in entities such as trees, rocks, rivers, roads, houses, bridges, etc. Our platoons that are represented by virtual human agents enter a specific area to perform a specified mission, which may be to attack, escape or just pass through a selected tactical area. The area contains static and/or moving platforms such as jeeps, planes, helicopters, commandos, and etc. The goal of the agents is to complete their mission in a group or by being divided into groups of two or more without being detected or caught by a platform that carries different kinds of sensors (Day TV, Infra-Red, SAR). The output views of the platform sensors are observed by the user at tactical command center in order to make the detection process realistic. Agents may follow rivers, go through the forests, and hide behind trees, run, or even crawl in order not to be seen. When any of the agents are detected and identified, they try to escape or hide to complete their mission until they are caught or terminated. 1