Integrating miniMin-HSP Agents in a Dynamic Simulation Framework

Lecture Notes in Computer Science, 2007

Proceedings of the fourth international conference on Autonomous agents, 2000

In this paper, we describe our current work to design and implement a system for rapid prototyping of 3D actors and integration of complex behavior modules. We identify the need and motivation for intelligent agent design tools for virtual reality environments and describe how our approach combines artificial intelligence technology and multi agent systems with 3D actors, while keeping consistence with three major objectives: 1) employment of a flexible development kit for the design of complex behavior for agents, 2)true rapid prototyping of 3D actors, including appearance, physical abilities and animations, and 3)the applicability of our system on general, arbitrary platform workstations. We present the CreatureViewer design system, and show how it supports the fast development of complex behavior by providing high level access to the specification of visual and abstract behavior of 3D actors. We also present CASA, a system for agent oriented design based on an extendible specification language and its according abstract interpreter model. Additionally, the entire system architecture of CreatureViewer is outlined, and finally, two examples are provided to show its performance and generic applicability.

Lecture Notes in Computer Science, 2003

In this paper, we propose the use of the Belief-Desire-Intention (BDI) model for cognitive agents for the implementation of animated characters. The BDI agent architecture has been widely used in dynamic and complex scenarios where agents may need to act under incomplete and incorrect information about other agents and the environment. In this work, we bring together an articulated model for character animation and an interpreter for AgentSpeak(L), an agent-oriented programming language that implements the BDI architecture. We have developed an interface that allows the BDI-based agent reasoning system to be used for guiding the behaviour of articulated characters in a virtual environment. This is a promising approach for the high-level specification of complex computer animations. The paper also presents a simple 3D animation that illustrates the use of BDI specifications in our approach.

Proceedings 1998 Australasian Computer Human Interaction Conference. OzCHI'98 (Cat. No.98EX234)

3D-graphics popularity has steadily increased i n a number of areas such as entertainment, scienti c visualization, simulation, and virtual reality. Despite this rapid growth the authoring of animated 3D objects in virtual environments is still by no means trivial. This article presents new concepts of the animated 3 D programming language SAM Solid Agents in Motion and its programming environment. In SAM, the main syntactic objects like agents, rules, and messages are represented as 3D objects. The design of a SAM program is supported b y a d e dicated 3D structure e ditor. The editor allows the de nition and spatial arrangement of SAM agents in a 3D scene by direct manipulation. The paper gives a number of SAM examples, demonstrating the authoring of simple animated virtual 3D scenarios. We thank Waldemar Rosenbach, Ralf Wegener and Rolf Zelder for implementing the SAM system and Volker Paelke and Sophie O'Halloran for proof reading.

In this paper we present experiments with search-based planning as the most generic description of intelligent behaviour for virtual actors. Using a MinMin heuristic search-based planner (HSP), we demonstrate how this can be used to generate efficient plans (in terms of plan length), and how the production of these minimum length plans is performed in times suitable for the real-time 3D virtual environments. Using an extension to the classical 'dinner-date' problem, we illustrate the planning and action execution undertaken by the virtual actor.

2005

In nearly any simulation system, there will be entities -that is, platforms or forces -that are not under the control of a human participant, either because the necessary personnel are not available or they would be too costly. These entities must mimic the behavior of the real-world entities that they represent in order to achieve some level of believability in the simulation. Crafting realistic, intelligent-seeming behaviors for simulated entities is a non-trivial task, however, and there are a variety of techniques that can be employed. While there exists a diverse array of agent architectures, two common approaches are cognitive architectures and state machines.

Proceedings of ACADIA 2002, 2002

Architectural design has relevance to the design of virtual worlds that create a sense of place through the metaphor of buildings, rooms, and inhabitable spaces. The design and implementation of virtual worlds has focused on the design of 3D form for fast rendering to allow real time exploration of the world. Using platforms that were originally designed for computer games, some virtual worlds now contain preprogrammed interactive behaviors. We present an agent model of virtual worlds in which the objects in the world have agency, that is, the objects can sense their environment, reason about their goals, and make changes to the environment. This agent model is presented and illustrated using a wall agent. Following from the wall agent, we generalize how agency can be attached to any 3D model in a virtual world.

2003

In this paper, we propose a new approach to virtual actors for 3D animation, in which the actors' behaviour can generate situations of narrative interest, as a consequence of the “limited rationality” of these actors. The actors behaviour is determined by plans generated in real-time in response to a single goal acting as a narrative drive. We use Heuristic Search Planning (HSP) as a non-optimal planning technique, which generate believable plans, yet susceptible of failure.

Many simulations of virtual environments involve realistically behaving autonomous agents to give users as much interaction capabilities as possible. In this kind of simulations, interoperable architectures such as the High Level Architecture enable us to create complex and lively environments from simple simulations of different kinds of entities. However, making simulations collaborate requires to give existing agents the ability to interact with the newly integrated ones. Such a task generally consists in redefining existing behaviours completely. Hence, making several even interoperable simulations collaborate, implies long and demanding developments while interoperability tries to prevent them. Such drawbacks are mainly due to the fact that most behavioural models are based on finite state machines and expert systems for which designers have to describe exhaustively all agents' behaviour. Classifier systems enable designers to describe agents' behaviours through goals ("what to do") instead of transitions or set of rules ("how to do it"). Then, agents learn how to reach those objectives using evolutionary learning algorithms. Such a modeling is more suited for the development of interoperable simulations. Indeed, adding new simulated agents only requires to add new goals to the existing ones. This paper presents the distributed driving simulation we built using interoperable subsimulations. At first, it presents how classifier systems make agents' behaviour easily evolutional in the context of interoperable simulations. Then, it shows how interoperability enables us to share the management of autonomous entities between several computers in a distributed way.

2001

Abstract In this paper, we describe a method for implementing intelligent behaviour for artificial actors in the context of interactive storytelling. We have developed a fully implemented prototype based on the Unreal™ game engine and carried experiments with a simple sitcom-like scenario. We discuss the central role of artificial actors in interactive storytelling and how real-time generation of their behaviour participates to the creation of a dynamic storyline.