Integrating miniMin-HSP Agents in a Dynamic Simulation Framework

Abstract

Abstract. In this paper, we describe the framework created for implementing AI-based animations for artificial actors in the context of IVE (Intelligent Virtual Environments). The minMin-HSP (Heuristic Search Planner) planner presented in [12] has been updated to deal with 3D dynamic simulation environments, using the sensory/actuator system fully implemented in UnrealTM and presented in [10]. Here, we show how we have integrated these systems to handle the necessary balance between the reactive and deliberative skills for 3D Intelligent Virtual Agents (3DI-VAs). We have carried out experiments in a multi-agent 3D blocks world, where 3DIVAs will have to interleave sensing, planning and execution to be able to adapt to the enviromental changes without forgetting their goals. Finally, we discuss how the HSP agents created are adequated to animate the intelligent behaviour of 3D simulation actors .

Simulating socially intelligent agents in semantic virtual environments

“…Planningbased agents are normally inspired by the STRIPS (ADL/PDDL) action language, where activities are represented using a classic state model with a conjunction of grounded literals. This approach has been used in storytelling domains (Charles et al, 2003) and other task-oriented 3D agents (Lozano et al, 2004). In these contexts, the usual representation of the actions suffers from different drawbacks when an actor is interested in reusing his/her operators with different objects.…”

Integrating Social Skills in Task-Oriented 3D IVA

“…Agents presented in [15] suffer from autonomous communication modules to coordinate their actuation or exchange information with their partners. Thus, the behavioral consistency of the multi-agent animations can be poor.…”

“…As previously exposed, characters are benevolent and their intention is to let the others finish their previous commitments. Following this premise, the miniMin-HSP planner used by our 3DIVA [15] will now start the search from a future virtual state resulted from applying the c beliefs over the current perceived state (this mental execution uses the add and delete lists of the STRIPS definition for the the external ongoing actions).In this manner, it constructs a prospective situation that skips actual dependences. Additionally, each c belief will lock the objects being used by other agents thereby forbidding to jeopardize the success of the tasks previously initiated (see figure 3).…”