No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
GameBots: A flexible test bed for multiagent team research
Abstract
The human factor in software development is the ingredient that ultimately gives a project team its soul.
... It thus is a suitable choice for putting an agent platform to the test. [8] argue that Unreal Tournament provides a useful testbed for the evaluation of agent technology and multi-agent research. These challenges also make UT2004 a suitable choice for dening a student project as students will be challenged as well to solve these problems using agent technology. ...
... Using the Unreal Tournament game as a starting point to connect an agent platform to thus does not limit possibilities to one particular game but rather is a rst step towards connecting an agent platform to a broad range of real-time environments. Moreover, a behavioral control layer called Pogamut extending Gamebots is available for UT2004 [10,8] which facilitates bridging the gap that exists when trying to implement an interface oriented towards high-level cognitive control of a game such as UT2004. ...
... Most projects that connect agents to UT2004 are built on top of Gamebots [8] or Pogamut [10], an extension of Gamebots: See e.g. [12,13] which use Gamebots and [7] which use Pogamut. ...
It remains a challenge with current state of the art technology to use BDI agents to control real-time, dynamic and complex environments. We report on our effort to connect the Goal agent programming language to the real-time game Unreal Tournament 2004. BDI agents provide an interesting alternative to control bots in a game such as Unreal Tournament to more reactive styles of controlling such bots. Establishing an interface between a language such as Goal and Unreal Tournament, however, poses many challenges. We focus in particular on the design of a suitable and reusable interface to manage agent-bot interaction and argue that the use of a recent toolkit for developing an agent-environment interface provides many advantages. We discuss various issues related to the abstraction level that fits an interface that connects high-level, logic-based BDI agents to a real-time environment, taking into account some of the performance issues.
... A third strand of work seeks to couple a mature agent platform with a particular fully-fledged ABM simulation. A number of special purpose ABM simulations have been developed that are designed to allow the integration of BDI agents, e.g., RoboCup [28], RoboCup Rescue [29], Gamebots [27], and the Multi-Agent Programming Contest [4]. Communication with the BDI platform is typically based on message passing (sockets), and so neutral with respect to the choice of BDI platform used to implement the agents. ...
... There has also been work on combining an ABM simulation with a particular cognitive agent platform [53][54][55]. Similar to the previous systems [27][28][29], these simulators support a single (parameterised) simulation model; however in addition they are limited to a single agent platform. For example, TacAir Soar [53] and SWARMM [55] are specific to a particular domain (air combat) and particular agent platform (Soar and dMARS respectively). ...
Agent-based models (ABMs) are increasingly being used for exploring and supporting decision making about social science scenarios involving modelling of human agents. However existing agent-based simulation platforms (e.g., SWARM, Repast) provide limited support for the simulation of more complex cognitive agents required by such scenarios. We present a framework that allows Belief-Desire-Intention (BDI) cognitive agents to be embedded in an ABM system. Architecturally, this means that the “brains” of an agent can be modelled in the BDI system in the usual way, while the “body” exists in the ABM system. The architecture is flexible in that the ABM can still have non-BDI agents in the simulation, and the BDI-side can have agents that do not have a physical counterpart (such as an organisation). The framework addresses a key integration challenge of coupling event-based BDI systems, with time-stepped ABM systems. Our framework is modular and supports integration of off-the-shelf BDI systems with off-the-shelf ABM systems. The framework is Open Source, and all integrations and applications are available for use by the modelling community.
... GameBots [Kaminka et al., 2002] is a virtual reality platform that allows users to create and evaluate agents in a 3D environment. GameBots is made up of two components. ...
... The authors developed a customised extension for the original Gamebots [Kaminka et al., 2002] testbed for connecting the Pogamut 3 toolkit to the UT2004 game. GameBots2004 exports game information through a TCP/IP text-based protocol that allows users to connect to the game through a client-server architecture. ...
Reinforcement learning (RL) is a paradigm which involves an agent interacting with an environment. The agent carries out actions in the environment and receives positive reinforcement for actions that are deemed “good” and penalties for “bad” actions based on a reward signal. The goal of the learning agent is to maximise the amount of reward it receives over time. This thesis presents several new behavioural architectures for controlling non-player characters (NPCs) in a modern first-person shooter (FPS) game using reinforcement learning. NPCs are computer-controlled players that are traditionally programmed with scripted, deterministic behaviours. We propose the use of reinforcement learning to enable the NPC to learn its own strategies and adapt them over time. We hypothesise that this will lead to greater variation in gameplay and produce less predictable NPCs. The first contribution of this thesis is the design, development and testing of two general purpose Deathmatch behavioural architectures called Sarsa-Bot and DRE-Bot. These architectures use reinforcement learning to control and adapt their behaviour. We demonstrated that they could learn to play competently and achieve good performance against fix-strategy scripted opponents. Our second contribution is the development of a reinforcement learning architecture, called RL-Shooter, specifically for the task of shooting. The opponent's movements are read in real-time and the agent chooses shooting actions based on those that caused the most damage to the opponent in the past. We carried out extensive experimentation that showed that the RL-Shooter architecture could produce varied gameplay, however, there was not a clear upward trend in performance over time. This led to our third contribution which involved developing extensions to the SARSA(λ) algorithm called Periodic Cluster-Weighted Rewarding and Persistent Action Selection. We designed these to improve the learning performance of RL-Shooter and we demonstrated that the use of the techniques resulted in a clear upward trend in the percentage hit accuracy achieved over time. Our final contribution is a skill-balancing mechanism that we developed, called Skilled Experience Catalogue, which is based on a by-product of the learning process. The agent systematically stores “snapshots” of what it has learned during the different stages of the learning process. These can then be loaded during the game in an attempt to closely match the abilities of the current opponent. We showed that the technique could successfully match the skill level of five different scripted opponents with varying difficulty settings.
... The development process made use of the Gamebots interface (Kaminka et al., 2002) and the PyPOSH implementation of BOD's Action Selection mechanism (Kwong, 2003), both of which are described and commented upon here. Improvements made to PyPOSH are also discussed. ...
... 25). Using PyPOSH with the Gamebots interface (Kaminka et al., 2002) allows agents in Unreal Tournament to be guided by POSH plans to perform actions coded in the Python programming language. A simple agent, "poshbot", was developed as part of PyPOSH's development, and the agent I have developed builds on this. ...
... The result allows agents to collaboratively maintain taskexecution conditions, and teamwork-structure conditions, throughout the execution of a task. To demonstrate DIESEL's contribution, we carried out experiments in the GameBots domain (Kaminka et al. 2002), evaluating the use of collaborativelymaintained maintenance conditions in contrast to existing approaches. ...
... To evaluate the contribution offered by DIESEL, we build a small two-agent team in the Game-Bots domain (Kaminka et al. 2002), an adversarial game environment that enables qualitative comparison of different control techniques (e.g., (D. Vu et al. 2003)). ...
There is significant interest in modeling teamwork in synthetic agents. In recent years, it has be- come widely accepted that it is possible to sepa- rate teamwork from taskwork, providing support for domain-independent teamwork at an architec- tural level, using teamwork models. However, ex- isting teamwork models (both in theory and prac- tice) focus almost exclusive on achievement goals, and ignore maintenance goals, where the value of a proposition is to be maintained over time. Such maintenance goals exist both in taskwork (i.e., agents take actions to maintain a condition while a task is executing), as well as in teamwork (i.e., agents take actions to maintain the team). This pa- per presents DIESEL, an implemented teamwork and taskwork architecture, built on top of Soar, that addresses maintenance goals in situated agent teams. We provide details of DIESEL's structure, and initial experiments demonstrating it in opera- tion in a dynamic rich domain.
... Still, numerous publications discuss game design as an internal part of research experiments. Irrespective of whether these are text-based MUDs (Slator et al. 1999), 3D game platforms (Lewis & Jacobson 2002, Bylund & Espinoza 2002, Kaminka et al. 2002 or game-like testing environments (Greenhalgh et al. 1999), they all apply game design to solve their research problems. One of the few exceptions of applying game design to non-game virtual environments is provided by Clarke-Willson (1998). ...
The focus of this thesis is on perceivable interaction forms that can be utilised in multi-player computer games and Collaborative Virtual Environments (CVEs). Interaction forms are manifestations of user-user and user-environment interaction. These forms convey the actions of the user to others, and to oneself, during synchronous communication and interaction partially resembling face-to-face encounters in the physical world.
The communication, co-ordination and collaboration difficulties in current 3D CVEs have yet to be overcome. The aim of this thesis is to conceptualise and delineate the available mutually perceivable interaction forms. The research analyses entertainment industry solutions and examines their potential as providers of design implications for CVE systems in general. Furthermore, games are used as platforms in the construction phases of the research. The research answers the problems of what interaction forms exist in multi-player games and what are their main characteristics and how to apply the understanding of possible interaction forms in order to help the design of games and professionally used CVEs? The problems are approached with conceptual analytical and constructive methods. Delineation and categorisation of various interaction forms in games and CVEs is presented. The gap between theoretical models and practical design is bridged by using several empirical cases as a constructive part of the research.
The significance of this research is related to the increasingly important role of CVEs in global organisations and virtual enterprises, as well as in the everyday life of people. The communicational and functional enhancements of the virtual environments and multi-player games make the interpersonal interaction more usable and applicable in computer mediated settings. The research provides a deeper understanding of the concept of interaction forms in the context of CVEs and multi-player games. The results can be used to create more supportive, appealing and communicative applications by applying interaction models and theories in the form of rich interaction design guidelines.
The main contributions of this research are the conceptual models of interaction forms and corresponding rich interaction design suggestions. The results reveal important and neglected aspects of interaction forms to the world of game and virtual environment designers. Combined with the descriptions of interaction design processes the overall outcome of the work is applicable to a wide audience designing or using computer systems that support communication, co-ordination and collaboration.
... In 2016 Gazebo was selected as simulation environment for the RoboCup Rescue Virtual Robot competition [24]. Previously, the RoboCup Rescue Virtual Robot competition was based on a simulation environment which was based on Gamebots on top of the Unreal Engine [12]. This simulation environment was called USARSim [3], which at the end also provided a ROS-interface [16]. ...
This review explores a natural learning curve which gives an appropriate RoboCup Rescue challenge at the right age. Children who got involved in the age group 14+ should continue their learning experience until they reach graduate level. To reduce the cost of such a learning experience, simulation is an attractive option in a large part of the world. The realism of the simulations and challenges should increase step-by-step, which are supported by more powerful but also more complex interfaces at each level/age-group. The result is a natural learning curve which allows for life-long learning. In this paper, we detail the requirements for such a platform and review a number of different simulation platforms and accompanying interfaces focusing on suitability for use for education rescue robotics. Resulting from this review of simulation platforms, a case-study of an example ‘game field’ rescue simulation platform suitable for students at different points along the learning curve.
... There are a number of general proposals for using games and game engines as test beds for multi-agent systems research in various ways (Kaminka et al., 2002;Bainbridge, 2007;Fabregues and Sierra, 2011;Koeman et al., 2018;Juliani et al., 2018), as well as more specific proposals arguing that role-playing games have useful design principles to offer (Barreteau et al., 2001;Guyot and Honiden, 2006). ...
Computer role-playing games (RPGs) often include a simulated morality system as a core design element. Games' morality systems can include both god's eye view aspects, in which certain actions are inherently judged by the simulated world to be good or evil, as well as social simulations, in which non-player characters (NPCs) react to judgments of the player's and each others' activities. Games with a larger amount of social simulation have clear affinities to multi-agent systems (MAS) research on artificial societies. They differ in a number of key respects, however, due to a mixture of pragmatic game-design considerations and their typically strong embeddedness in narrative arcs, resulting in many important aspects of moral systems being represented using explicitly scripted scenarios rather than through agent-based simulations. In this position paper, we argue that these similarities and differences make RPGs a promising challenge domain for MAS research, highlighting features such as moral dilemmas situated in more organic settings than seen in game-theoretic models of social dilemmas, and heterogeneous representations of morality that use both moral calculus systems and social simulation. We illustrate some possible approaches using a case study of the morality systems in the game The Elder Scrolls IV: Oblivion.
... Moreover, it has many services including multiplayer capability, more detailed graphics, and flexible plugin architecture [22]. ...
Virtual Reality (VR) technology has become one of the most advanced techniques that is used currently in many fields. The role of education is extremely important in every society; therefore, it should always be updated to be in line with new technologies and lifestyles. Applying technology in education enhances the way of teaching and learning. This paper clarifies a virtual reality application for educational resolutions. The application demonstrates a virtual educational environment that is seen through a Virtual Reality headset, and it is controlled by a motion controller. It allows the user to perform scientific experiments, attend online live 360° lectures, watch pre-recorded lectures, have a campus tour, and visit informative labs virtually. The application helps to overcome many educational issues including hazardous experiments, lack of equipment, and limited mobility of students with special needs.
... Dashed arrows represent data flow, and solid arrows represent control. (Ontañón et al. 2013) The design of game-bots (Kaminka et al. 2002) (in Figure 2.7) solve problems in playing StarCraft by decomposing problems into sub-problems where each of them focuses on a specific aspect of gameplay. By analyzing the architecture of the bots, they put forward two main ideas behind the design. ...
By nature, human beings are curious about their environment. Arriving in
a new place, they observe, recognize and interact with their surroundings. People collect information about the new place, and locate objects in that space that help them to make further decisions. This is a typical scenario of spatial exploration. Spatial exploration is common human behavior, where humans explore unknown environments to acquire information and resources. It is pervasively seen in real-world and virtual environments, from exploring new living/working spaces to charting the oceans or venturing beyond the boundaries of our planet. Just as humans explore ‘real’ environments, they also investigate artificial environments in video games. Computer agents, which perceive surrounding environments with limited visual range, often appear in exploration activities, acting as tools or partners for explorers. Despite the broad range of human activities that employ exploration behavior, this element has been insufficiently investigated and understood. Additionally, even though it is commonly accepted that believable agents benefit people in human-computer interaction systems, the research into creating computer agents with believable exploration behavior has been neglected. To solve these issues, I extract the patterns of human exploration behavior in virtual environments, and explore the methodologies of developing believable agents, which explore spatial environments in human-like ways. In the pursuit of this goal, this thesis
makes the following four contributions to the emerging field of believable agent exploration: 1) I employed video games as a testbed to investigate human behavior of spatial exploration. Human players played specialized exploration games, verbalized their behavior during playing and discussed their thoughts in the post-play interview. Behavioral patterns were extracted based on replays of playing, think-aloud data and interview data via thematic analysis. 2) Differences of exploration behavior between human and computer agents were identified through a third person-observation assessment of believability. 3) A heuristic agent was developed, which mimics human exploration methods reflected via the behavioral patterns. Three heuristics, as components of the heuristic agent, were designed to filter potential options when the agent decides where to explore in each step. 4) An integrated agent was developed by filling the behavior gaps between human and computer agents, where an integrated architecture embedded expectations of humanlike
exploration from mid-level players. Both the heuristic agent and the integrated agent passed the third-person-observation assessment of believability. Therefore, findings in this thesis contribute to fill the gaps in the fields of understanding human exploration behavior as well as developing believable agent.
... Agent automata [302,303] can ascribe human agency to agent-characters [302,304,305] or functional agency to urban forms [306]. Multi-agent systems [307] are useful when assigning different motives and behaviors to classes of agents based on their role in streetscape dynamics [308,309]. Cellular automata [310] are commonly employed to tessellate streets into automata units and transition the flow of individuals or crowds across the lattice that they form in streetscape models [138,311,312]. Geographic automata [313] make use of GIS to flexibly specify movement and neighborhood relations [65] across the built and human substrate of synthetic streetscape phenomena. ...
Streetscapes have presented a long-standing interest in many fields. Recently, there has been a resurgence of attention on streetscape issues, catalyzed in large part by computing. Because of computing, there is more understanding, vistas, data, and analysis of and on streetscape phenomena than ever before. This diversity of lenses trained on streetscapes permits us to address long-standing questions, such as how people use information while mobile, how interactions with people and things occur on streets, how we might safeguard crowds, how we can design services to assist pedestrians, and how we could better support special populations as they traverse cities. Amid each of these avenues of inquiry, computing is facilitating new ways of posing these questions, particularly by expanding the scope of what-if exploration that is possible. With assistance from computing, consideration of streetscapes now reaches across scales, from the neurological interactions that form among place cells in the brain up to informatics that afford real-time views of activity over whole urban spaces. For some streetscape phenomena, computing allows us to build realistic but synthetic facsimiles in computation, which can function as artificial laboratories for testing ideas. In this paper, I review the domain science for studying streetscapes from vantages in physics, urban studies, animation and the visual arts, psychology, biology, and behavioral geography. I also review the computational developments shaping streetscape science, with particular emphasis on modeling and simulation as informed by data acquisition and generation, data models, path-planning heuristics, artificial intelligence for navigation and way-finding, timing, synthetic vision, steering routines, kinematics, and geometrical treatment of collision detection and avoidance. I also discuss the implications that the advances in computing streetscapes might have on emerging developments in cyber-physical systems and new developments in urban computing and mobile computing.
... The Unified System for Automation and Robot Simulation (USARSim) environment has been used for many years by robotics researchers and developers as a validated framework for simulation [6,3]. The original version was developed in 2003, based on the concept of GameBots [10]. ...
Since the first demonstration of the Virtual Robot Competition, USARSim has been used as the simulation interface and environment. The underlying simulation platform, Unreal Engine, has seen three major upgrades (UT2004, UT3 and UDK). These upgrades required a whole new USARSim simulator to be built from scratch. Yet, between those versions the USARSim interface has not been modified, which made USARSim a stable platform for more than 10 years. This stability allowed developers to concentrate on their control and perception algorithms. This paper describes a new prototype of the USARSim interface; implemented as plugin to Gazebo, the simulation environment native to ROS. This plugin would facilitate a shift of the maintenance of the simulation environment to the Open Source Robotics foundation and attract new teams to the Virtual Robot Competition.
... We have developed an adapted version of the CERA-CRANIUM cognitive architecture with the aim to explore the application of MC implementations to the domain of multiplayer video games. We have used GameBots [41] and the Pogamut 2 platform [42] to interface with the UT2004 game server (as CERA-CRANIUM is implemented in .NET Framework and Pogamut 2 is written in the JAVA programming language, we have used IKVM [43] to generate a fully working .NET version of the Pogamut 2 Core library). Fig. 6 depicts the experimentation environment in which the tests have been performed. ...
This chapter introduces Machine Consciousness as a new research field applied to the development of human-like behaviour of non-player characters (NPCs) in video games. Key aspects, advantages, and challenges of this young research area are discussed using the cognitive architecture CERA-CRANIUM as an illustrative example of an autonomous control system inspired by cognitive theories of human consciousness. Additionally, other cognitive architectures used in video games are also analyzed. The bot codenamed CC-Bot2, winner of the 2K BotPrize 2010 competition and based on the CERA-CRANIUM cognitive architecture, is also described in this chapter. Specifically, the particular way in which CC-Bot2 processes the sensory-motor information and generates sequences of adaptive human-like actions is discussed. We also analyze the main differences between CC-Bot2 and other bots, focusing on the key features that allowed CC-Bot2 to win first place in the competition. Finally, we conclude by describing the main lines of work for future CC-Bot implementations and pointing out major conclusions about the application of Machine Consciousness to the design of believable bots. © 2012 Springer-Verlag Berlin Heidelberg. All rights are reserved.
... The main objective of Pogamut 3 is to simplify the coding of actions taken in the environment, such as path finding, by providing a modular development platform. The toolkit integrates the Unreal Tournament 2004 game, GameBots2004 (Kaminka et al. 2002), the GaviaLib library (Pogamut Project Core), the Pogamut agent and the NetBeans IDE 6 . A detailed technical description can be found in Gemrot et al. (2009). ...
As graphics in modern computer games move closer to photorealism, the emphasis for game developers is switching towards improving the in-game Artificial Intelligence (AI). Traditional scripting and rule-based systems are being replaced by more intelligent and immer-sive approaches. The goal of AI in computer games is to create intelligent autonomous agents that mimic human behaviour as closely as possible, in order to create a challenging yet enjoyable experience for human players. This paper describes the application of Reinforcement Learning (RL), an approach inspired by how humans learn, to the creation of intelligent "bots " in a First Person Shooter (FPS) game.
... In an attempt to overcome this limitation some research work have proposed the use of agent centered approaches to control characters. One of the first attempts to connect intelligent agents to games was Gamebots [23]. It provided an infrastructure to allow the connection of any agent platform and the Unreal Tournament game. ...
Looking at the ways in which players interact with computer games (the gameplays), we perceive predominance of character-centered and/or microcontrolled modes of interaction. Despite being well established, these gameplays tend to structure the games in terms of challenges to be fulfilled on an individual basis, or by thinking collectively but having to microcontrol several characters at the same time. From this observation, the paper presents a complementary gameplay in which the player is urged to face collective challenges by designing character organizations. The basic idea is to make the player structure and control group of characters by defining organizational specifications (i.e., definitions of roles, collective strategies, and social norms). During the game, commanded by the player, artificial agents are then instantiated to play the roles and to follow the strategies and norms as defined in the organizational specification. To turn the idea into practice, the paper proposes an abstract architecture comprising three components or layers. This architecture is materialized in a proof of concept prototype that combines the Minecraft game server, JADE agent platform, and MOISE+ organizational model. Variations and possibilities are discussed and the proposal is compared to related work in the literature.
... GameBots is a modification to the Unreal Tournament game that allows bots to be controlled through a normal TCP/IP socket [7]. With a simple text-based TCP/IP protocol GameBots can be used to create and manipulate bots in an Unreal Tournament instance. ...
Urban Search and Rescue (USAR), is a difficult, often ill-defined problem involving coordination between robots, agents, and people to execute a multitude of tasks and adapt to complex environmental challenges. Based on standards developed by the National Institute of Standards and Technology (NIST) USAR Test Facility, we implemented a variety of two-wheeled differential drive robots, built several user interfaces for the robots, and developed a game engine-based multi-robot simulation. In this paper we describe our approach in which user interface, control, and coordination strategies can be evaluated in simulation and then moved directly to testing in prototyped robots in a real arena; and illustrate an example of a robot system designed and implemented under this model. The interoperability of code between the simulated and physical robot system significantly shortens the cycle between concept and test.
... The classical teambased game type Capture The Flag was chosen to evaluate the team coherence, because it is easy to measure the results and simple to identify the team behavior. The GameBOTs [Andrew Scholer 2000] [Kaminka et al. 2002] project is used to communicate the developed bot with the game. ...
The demand for believable behavior in computer games has driven the research in artificial intelligence to improve character's behavior in games. However, little has been done to improve the collective behavior in partially observable, dynamic, and stochastic environ-ments. In this work, teams of agents are developed, based on Joint Intentions, for environments with such features as the game Unreal Tournament 2004. Because of some limitations of existing tools, we introduce a new proxy-based tool to help agents to be a team-mate. Experiments have shown that it is feasible to use a distributed approach and the advantages of our tool face to Machinetta.
... Research has been done over the last decade when it comes to creating agent control mechanisms for VW. The first project of note is Gamebots (Kaminka et al., 2002), which was a client-side approach to controlling Unreal Tournament 2 2 http://www.unrealtechnology.com/technology.php , an engine used for first-person shooter games. ...
Incorporating autonomy and intelligence into virtual worlds to build an engaging virtual environment for applications such as serious games is becoming more desirable. There are challenges in integrating these systems including concerns with synchronization, communication, monitoring, efficiency, and control. This paper presents an approach to integrating a virtual world engine with a multiagent system platform through the creation of an interface between them. We show the feasibility and effectiveness of the approach through developing agents controlled non-player controlled characters (NPCs) in Open Wonderland and pur-poseful communication channels among agents using Jason AgentSpeak.
... Os experimentos foram executados com o jogo Unreal Tournament 2004, gravando as partidas através da interface GameBots [Kaminka et al. 2002]. O ciclo de aprendizado é realizado por partidas, onde ao final de cada uma, os logs são analisados a fim de atualizar o modelo do ambiente e para identificar como os personagens se comportam. ...
Resumo A demanda por comportamentos convincentes em jogos de com-putador tem orientado a pesquisa em inteligência artificial com o objetivo de aprimorar o comportamento dos personagens em jogos. A fim de alcançar um comportamento convincente, os personagens devem usar o conhecimento sobre o mundo no qual atuam. A repre-sentação do ambiente é muito importante para executar um compor-tamento mais sofisticado, mas na maioria dos jogos essa represen-tação ainda continua sendo definida manualmente de forma estática pelo projetista do ambiente, levando os personagens a um compor-tamento previsível e pouco convincente. Apesar da geração auto-mática de posições táticas estar sendo tratada por alguns trabalhos, não existe uma solução geral e as abordagens atuais possuem mui-tas limitações. Em função disto, este trabalho apresenta alguns re-sultados preliminares do uso de agrupamento espaço-temporal para encontrar posições táticas em ambientes virtuais, permitindo que os bots se adaptem ao modo como as partidas são jogadas. 1 Introdução Na maioria dos jogos, informações táticas 1 são especificadas no ambiente pelo projetista do mundo virtual. Entretanto, a especifica-ção dessas regiões interessantes durante a criação dos mapas é um trabalho que consome muito tempo. Em geral, essa tarefa é reali-zada manualmente, o que pode limitar as interações no ambiente, dado que algumas localizações interessantes podem não ser identi-ficadas pelo projetista. Algumas sequências de ações simples, como ações lineares prede-finidas, podem ser orquestradas com sucesso pelo projetista do am-biente através da especificação de posições táticas. Entretanto, seu trabalho leva em consideração apenas o ambiente estático, e não avalia as interações entre personagens no ambiente. Além disso, a especificação manual de posições táticas pode ser um trabalho pesado, sendo suscetível a falhas, dado que o projetista pode não prever todos os pontos interessantes que podem surgir no ambiente durante as partidas.
... Research has been done over the last decade when it comes to creating agent control mechanisms for VW. The first project of note is Gamebots (Kaminka et al., 2002), which was a client-side approach to controlling Unreal Tournament 2 2 http://www.unrealtechnology.com/technology.php , an engine used for first-person shooter games. ...
Incorporating intelligence and social behaviours into virtual worlds for learning is becoming more desirable in making them smart, adaptive, personalized, and therefore, more effective and engaging. Realistic non-player controlled characters (NPCs) are essential of a game world and are making the virtual world more real for players. This is true in video games where more interactive NPCs support the story narrative of a game, making the game more immersive, more convincing, but it is also true in other areas where virtual worlds are used such as business and education, increasing the effectiveness of those environments. Research that has been done with virtual agents and multi-agent systems can be leveraged to create more realistic NPCs through purposeful communication channels, inter-agent interactions and environment-agent interactions for game-based learning applications. This research proposes an approach to controlling avatars with intelligent agents through the creation of an interface between a multi-agent system to a virtual world engine. Basic NPC behaviours controlled by agents using Jason Agent Speak are used to test the feasibility of the approach. A Quizmaster is designed and illustrated as a proof of concept of the use of such agent controlled avatars in educational context.
... The team goal is to fill bins placed in environment to match the specified pattern as fast as possible. A 3D continuous dynamic environment forcreating and evaluating intelligent agents is introduced in [8]. This test bed not only has a set of predefined tasks and environments but also allows extending them in various ways. ...
One of the most interesting domains in multiagent systems is teamwork between a group of agents and related issues. In this paper we introduce a multi context test environment for simulating real world uncertainty and dynamicity in agents' teamwork. This environment is a new complex version of famous wumpus world with some features changed to be tailored for real and complex situation. We have implemented it in java language with using of JADE platform so that it can be applied for teamwork issues easily. This application provides high flexibility and easy to use interface so that users can make use of it conveniently. We believe this environment can be used for evaluation of any subjects related to multi-agent's joint operation such as cooperation, coordination, negotiation, planning, mental model sharing and etc.
... Shown in Figure 22 are the real and simulated NIST reference arenas [49]. Finally, USARSim applies a modification of GameBots [52] to communicate with the virtual robot via a network, which makes the simulator independent of both the programming language and the computer platform. Moreover, USARSim provides tool kits to help users control robots via popular software like Player and Pyro. ...
Abstract Advances in robotic technologies and artificial intelligence are allowing robots to emerge from research laboratories into our lives. However, experience in field applications such as homeland security, search and rescue, health care, personal assistance, entertainment etc. isshowing that we have underestimated the importance,of human-robot interaction (HRI) and that new problems,arise in HRI as robotic technologies expand. HRI is an active research area that is still in its infancy and has not yet developed,clear classificationsof topics or metrics. This thesis classifies HRI along four dimensions -- human, robot,task and world and illustrates that previous HRI classifications can be successfully interpreted as either about one of these elements or about the relationship between,two or more of these elements.
... We achieved flexibility and low cost in the game environment by following in the footsteps of other projects (MIMESIS (Young et al. 2004), ESC Online (Martin 2001), and Gamebots (Kaminka et al. 2002)) by using the Unreal Tournament (UT) engine. UT provides an affordable, off-the-shelf 3D game engine that can be easily extended. ...
... The practicality of using computer games has increased as many game makers have begun providing better software interfaces to their products [Lewis and Jacobson, 2002]. The efficacy and legitimacy of using computer games for AI research is evident by the increasing number of researchers using games such as Unreal Tournament, Quake and Half-Life as simulators [Lewis and Jacobson, 2002;Laird, 2001;Bylund and Espinoza, 2002;Jacobson and Hwang, 2002;Kaminka et al., 2002;Piekarski and Thomas, 2002]. Laird [2001Laird [ , 2000; Laird and Lent [2001] cite several additional advantages to using computer games as an AI test-bed. ...
Abstract Robotic agents in dynamic environments must regularly navigate reactively, relying solely on their immediate local perceptions to make navigation decisions. In environments with complex topography, features such as terrain undulation, geometrically complex barriers and similar obstacles form local maxima,and minima that can trap and hinder agents navigating reactively. When agents are operating in groups, navigation is even more dicult, requiring agents to explicitly communicate to plan and coordinate their actions. The cost of explicit communication can be substantial, however, making it desirable to avoid its use in many domains. Accordingly, in this thesis I present methods to allow a team of agents using reactive navigation to assist one another in their explorations through implicit communication. Agents accomplish this by modifying their environment in a manner meaningful to other agents in order to share knowledge about advantageous locations. The eectiveness,of the various methods presented is shown in a number of experiments in which a team of agents is required to locate a goal situated somewhere
... To directly monitor the performance of human players, we customized the first-person shooter game, Unreal Tournament (UT), using the game development language Unrealscript. Many of the original UT game classes were written in Unrealscript and thus can be directly subclassed to produce modified versions of the game (known as mods); for example, Gamebots [20] is an example of a mod that allows external programs to control game characters using network sockets. ...
Abstract This thesis focuses on the problem of activity recognition for physically-
... Still, numerous publications discuss game design as an internal part of research experiments. Irrespective of whether these are text-based MUDs (Slator et al. 1999), 3D game platforms (Lewis & Jacobson 2002, Bylund & Espinoza 2002, Kaminka et al. 2002 or game-like testing environments (Greenhalgh et al. 1999), they all apply game design to solve their research problems. One of the few exceptions of applying game design to non-game virtual environments is provided by Clarke-Willson (1998). ...
... The system can manage multiple agents simultaneously and can be easily plugged into another virtual environment, provided only with a gamebots-like interface (API) to that environment. In academia, UT with Gamebots have been used in several research projects already [18,23], which makes it valuable for comparing AI approaches from different laboratories. The goal of Pogamut is to extend it to create a platform that can be used extensively by students and new programmers. ...
Agent based systems are becoming popular outside the agents research community, among biologist, artists and in the game industry. Yet tools are lacking that facilitate non-expert agent developers building complicated agents for modelling and systems. As a part of our own agent-based research programmes we have been developing such tools. In this paper, we review the progress made, highlighting issues of usability. Examples of agents developed in these tools are also given, with an emphasis on intelligent virtual combat agents situated in Unreal Tournament.
... The simulated world was run through an Unreal Tournament 2004 server program. An Unreal Tournament 2004 modification called Gamebots, developed and maintained by a separate party, was used to open the server program to any TCP/IP client (Kaminka 2002). Another Unreal Tournament 2004 modification call USARSim, Urban Search and Rescue Simulator, also developed and maintained by a separate party, extends the Gamebots modification to support robot related messages (Wang et al.). ...
... Relatively to the use of game engines for virtual simulations, the literature reports their use for other applications besides entertainment (Zyda, 2007;Carpin et al., 2007;Jacobson and Lewis, 2005;Kot et al., 2005;Rosenbloom, 2003;Wang et al., 2003;Zyda et al., 2003;Swartout and van Lent, 2003;Manojlovich et al., 2003;Tsang et al., 2003;Rhyne, 2002;Lewis and Jacobson, 2002;Laird, 2002;Piekarski and Thomas, 2002;Jacobson and Hwang, 2002;Kaminka et al., 2002;Bylund and Espinoza, 2002). Due to the high costs involved in the development of totally new platforms for virtual simulations, game engines bring the kind of platform that fulfills researchers' needs. ...
This work reports results obtained with the use of a game engine as a tool to create and to navigate in virtual environments, to perform simulations and training of workers in risky areas, for safety purposes. The game engine used can be used for non-commercial and educational purposes, and its source code is available for users. Thus, the engine can be modified and adapted for the modeling and simulation of any environment, including the development of new functionalities. A case study is shown, which has the purpose of supporting dose assessment in nuclear plants, for optimization of operational routines in these areas. The implemented modifications are explained, and simulations' results are shown.
... The virtual simulation environments must provide the facilities and interfaces to allow for the developer to create new worlds, provide virtual sensing capabilities, modify objects, and appropriate command and control channels to interface with all agents in both the real and virtual environments [33], [34]. The virtual world must also provide the facilities to log vital information in real-time and visualization tools to access and display the internal state of the vehicle during operation. ...
The development of performance metrics is critical in the evaluation and advancement of intelligent systems. Obtaining the pinnacle of intelligence in autonomous vehicles requires evolutionary standards and community support. In order to analyze and compare competing implementations of intelligent systems, the critical components of the system must be decoupled to facilitate repeatable trials that target specific aspects of the system's overall task. This paper presents a framework for a real virtual simulation environment that provides the facilities and tools to formally test the limitations and capabilities of autonomous road driving vehicles.
... Because contemporary games are designed so that users can modify both the environment (level programming) and behavior (game programming) of the game, they make powerful, inexpensive tools (Lewis & Jacobson 2002) for researchers who need interactive 3D modeling and graphics. Epic Game's Unreal 2003 game engine used in Unreal Tournament 2003 and America's Army (Zyda et al., 2003) among other games is especially attractive because of the Gamebots (Kaminka et al., 2002) modification written to allow entities within the game to be controlled through standard TCP/IP sockets. Using the Gamebots interface, data from within the simulation such as ranging data showing the robot's distance from walls and objects can be read off the socket while commands are sent back to control the robot within the simulation. ...
We are developing simulations of the National Institute of Standards and Technology (NIST) Reference Test Facility for Autonomous M obile Robots (Urban Search and Rescue) in order to develop and test our strategies for Robots-Agents-People (RAP) team coordination and control. The NIST USAR Test Facility is a standardized disaster environment consisting of three scenarios of progressive difficulty: Yellow, Orange, and Red arenas. The USAR task focuses on robot behaviors, and physical interaction with standardized but disorderly rubble filled environments. As part of our research effort we are constructing and permanently housing a physical replica of the Orange arena at Carnegie Mellon University. A simulation of the Orange arena was constructed first i n order to allow comparisons between simulated and real environments as s oon as construction of the physical Orange Arena is completed. We hope to use the simulations to provide a testbed in which to evaluate rapidly prototyped interfaces and control strategies prior to the c onstruction and testing of physical robots. This paper describes our simulation approach based on the use of the Unreal game engine to provide graphics and physics and simplified CAD models textured from digital photographs to model the environment.
... Various modifications to the UT engine have been made in recent years to enhance its support for different applications. In particular, Kaminka and et al. developed the Gamebots as an infrastructure for multi-agent research [10]. Sentry Studios developed a popular UT modification, Infiltration, which replaced the futuristic players/bots in UT with more realistic soldier and weapon models [11]. ...
We describe our work on Twilight City, a virtual training environment for Military Operations on Urbanized Terrain (MOUT). MOUT are often characterized by building-to-building, room-to-room, and person-to-person close combat, therefore the virtual training environment for MOUT requires high fidelity in 3D graphics and realistic behaviors of various objects. Twilight City is constructed on a commercial multi-player first-person shooter (FPS) game engine, the Unreal Tournament (UT). Various modifications to the UT game engine have been made to incorporate various aspects in real MOUT scenarios. Case studies show that the AI bots are able to demonstrate some human-like tactical behavior. Human factor tests are conducted to evaluate the performance of Twilight City. The results demonstrate the effectiveness of Twilight City as an urban warfare simulation environment. The lessons learned from these experiments are also discussed which provide directions for future improvement.
... The developed navigation algorithms were exercised, tested and verified using a simulation infrastructure comprised of USARSim [2,4,15], a Gamebots [1,6] variant of the Unreal Engine [14] and the Mobility Open Architecture Simulation and Tools (MOAST) [10] framework [12]. Using an arbitrary navigation path of multiple segments (Fig. 7) the algorithms were executed using a simulated, Ackerman-steered, automated guided vehicle (AGV) to navigate the path. ...
Autonomous robot navigation is one of the most important areas of robotics research. While waypoint-based navigation is well understood, particular design characteristics (e.g. steering mechanisms) of a robotic platform may render general navigation approaches inaccurate or ineffectual. In this paper, we develop a novel, deliberative-planning, waypoint-based robot navigation algorithm that allows accurate path following for Ackerman-steered vehicles. The algorithm is validated via simulation. Collected performance data provide insight on the relation of the algorithm to other mobility factors such as velocity.
Simulation environments have always been a vital part of robotics research, providing tools for modeling and testing novel concepts and algorithms. In the early years, robotic simulators were restricted to 2D environments, without being fully able to simulate the complexity of the real world. Both the progress of computer hardware technologies and the advances in the computer graphics domain, enabled the development of visually realistic and physically accurate simulators. The gradual improvement on the performance of robot perception, cognition, and decision-making algorithms had a great impact on the way that robotic systems understand and interact with their environments, enabling them to even operate autonomously. The demand to simulate such complex robotic systems and the need to collect diverse, large-scale, and realistic data to train and test deep learning algorithms before being deployed on real-world scenarios, pushed modern simulation environments to even higher standards. The aim of this chapter is to present a systematic review of robotic simulation environments, while also presenting their evolution over the years.
The paper summarizes a number of recent application domains (ADs) for large-scale multiagent systems (LSMASs) including the Industry 4.0, cooperative intelligent transport systems (C-ITSs), smart grids, smart cities, virtual & augmented reality (VAR) and electronic sports (eSports). These ADs intentionally overlap to show some of the basic building blocks of modern large scale applications-organization , cooperation, competition, robustness and (col-lective) intelligence. Benefits of using organization-centred LSMASs are outlined and a discussion is provided .
Dans les pays développés tels que le Canada ou la France, la population est vieillissante et le nombre de personnes atteintes de déficiences cognitives augmente en conséquence. Ces troubles ont des conséquences sur les activités de la vie quotidienne pour les personnes qui en souffrent. Selon l’autonomie de ces personnes et la sévérité de leur déficience, un hébergement en centre spécialisé peut être envisagé. Ces centres spécialisés représentent souvent un coût financier énorme tant pour la personne que pour la société. Afin de limiter ces coûts, une solution alternative a émergé : les habitats domotiques. Ce sont des habitats dans lesquels un ensemble de technologies permet de pallier aux déficiences des personnes et de leur donner une autonomie accrue. L’activité de préparation d’un repas est une activité complexe qui peut présenter des risques variés pour des personnes atteintes de déficiences. Ces personnes vivent rarement seules, il faut tenir compte qu’un public varié puisse bénéficier du système. Ces habitats sont généralement déjà équipés d’appareils, il devient nécessaire pour le système de pouvoir s’adapter à ces appareils existants. L’objectif de ces travaux est la réalisation d’un prototype permettant d’assurer la sécurité lors de l’activité de la préparation d’un repas par des personnes atteintes de la maladie d’Alzheimer et ses aidants (professionnels ou naturels). Ce prototype doit s’adapter au profil des usagers, de son environnement et du matériel sur lequel il est déployé. Pour ce faire, le système, basé sur un système multi agent, applique des règles de sécurité qui se personnalisent par le biais du profil médical des usagers. Nos travaux ont été menés au sein de deux laboratoires, qui chacun, disposent d’appareils de cuisine différents dans leurs habitats intelligents pour la santé. Le système a pu être testé dans ces deux environnements, son adaptation vis-à-vis d’une clientèle variée et pour plusieurs risques de sécurité à travers des scénarios d’usage. Les résultats de ces expérimentations ont permis de montrer que le prototype répond bien aux objectifs visés.
In this paper we propose a methodology for immersive multiprojection visualization of manufacturing processes. It admits sceneries with dynamic components and allows Virtual Reality collaborative visualization among geographically distributed users, through CAVE devices. We use a set of tools for modeling, converting, visualizing and interacting, besides an game engine as renderization core system. The method can be applied to CAD projects, models and simulations used in industry. The proposed ideas are then validated through the study of a real case related to the Shipbuilding and Offshore Industries. Digital Manufacturing Shipbuilding Virtual Reality
One of the key success factors that contribute towards the creation and sustenance of online (2D and 3D) virtual worlds for learning might be to provide game-style educational activities. However, there is no development platform available which can meet the inherent system requirements including usability of platform and scalability to modern massively multiplayer online games, yet focused on engaging learning for the individual user. Work has been done with software agents in the context of multiagent systems (MAS), and it makes sense to try to leverage that work when it comes to modeling functional modules, controlling realistic non-player characters (NPCs), and Personal Assistants for Learning (PALs) in a virtual learning world. There are challenges to integrating a multi-agent system into a virtual world including concerns with synchronization, communication, monitoring, efficiency, and control. This paper describes the design and implementation of an integrated architecture for performing and facilitating quiz games for adaptive testing with a multi-agent system JADE/Jason and a 3D virtual world engine Open Wonderland.
Computer game industry is one of the most profitable nowadays. Although this industry has evolved fast in the last years in different fields, Artificial Intelligence (AI) seems to be stuck. Many games still make use of simple state machines to simulate AI. New models can be designed and proposed to replace this jurassic technique. In this paper we propose the use of Artificial Neural Networks (ANN) as a new model. ANN will be then in charge of receiving information from the game (sensors) and carry out actions (actuators) according to the information received. The search for the best ANN is a complex task that strongly affects the task performance while often requiring a high computational time. In this work, we present ADANN, a system for the automatic evolution and adaptation of artificial neural networks based on evolutionary ANN (EANN). This approach use Genetic Algorithm (GA) that evolves fully connected Artificial Neural Network. The testing game is called Unreal Tournament 2004. The new agent obtained has been put to the test jointly with CCBot3, the winner of BotPrize 2010 competition [1], and have showed a significant improvement in the humannesss ratio. Additionally, we have confronted our approach and CCBot3 (winner of BotPrize competition in 2010) to First-person believability assessment (BotPrize original judging protocol), demonstrating that the active involvement of the judge has a great impact in the recognition of human-like behaviour.
Imitation is a powerful and pervasive primitive underlying examples of intelligent behaviour in nature. Can we use it as a tool to help build artificial agents that behave like humans do? This question is studied in the context of the BotPrize competition, a Turing-like test where computer game bots compete by attempting to fool human judges into thinking they are just another human player. One problem faced by such bots is that of human-like navigation within the virtual world. This chapter describes the Human Trace Controller, a component of the bot which took second place in the BotPrize 2010 competition. The controller uses a database of recorded human games in order to quickly retrieve and play back relevant segments of human navigation behaviour. Empirical evidence suggests that the method of direct imitation allows the bot to effectively solve several navigation problems while moving in a human-like fashion. © 2012 Springer-Verlag Berlin Heidelberg. All rights are reserved.
Games are good test-beds to evaluate AI methodologies. In recent years, there has been a vast amount of research dealing with real-time computer games other than the traditional board games or card games. This paper illustrates how we create agents by employing FALCON, a self-organizing neural network that performs reinforcement learning, to play a well-known first-person shooter computer game called Unreal Tournament. Rewards used for learning are either obtained from the game environment or estimated using the temporal difference learning scheme. In this way, the agents are able to acquire proper strategies and discover the effectiveness of different weapons without any guidance or intervention. The experimental results show that our agents learn effectively and appropriately from scratch while playing the game in real-time. Moreover, with the previously learned knowledge retained, our agent is able to adapt to a different opponent in a different map within a relatively short period of time.
This paper proposes a methodology for the distributed visualization of Manufacturing Processes. The methodology consider the use of a group of tools capable to propitiate the immersive visualization of virtual sceneries, in which CAD projects and simulations obtained from manufacturing tools are used in the industry. The proposal is validated in a real case study associated of Shipbuilding and Offshore Industries.
This chapter describes initial research into intelligent agents using the Beliefs-Desires-Intentions (BDI) architecture in a human-machine teaming environment. The potential for teaming applications of intelligent agent technologies based on cognitive principles will be examined
One issue confounding the development and experimentation of distributed modeling and simulation environments is the inability of the project team to identify and collaborate with resources, both human and technical, from outside the United States. This limitation is especially significant within the human behavior representation area where areas such as cultural effects research and joint command team behavior modeling require the participation of various cultural and national representatives. To address this limitation, as well as other human behavior representation research issues, NATO Research and Technology Organization initiated a project to develop a NATO virtual institute that enables more effective and more collaborative research into human behavior representation. However, in building and operating a virtual institute one of the chief concerns must be the cyber security of the institute. Because the institute "exists" in cyberspace, all of its activities are susceptible to cyberattacks, subterfuge, denial of service and all of the vulnerabilities that networked computers must face. In our opinion, for the concept of virtual institutes to be successful and useful, their operations and services must be protected from the threats in the cyber environment. A key to developing the required protection is the development and promulgation of standards for cyber security. In this paper, we discuss the types of cyber standards that are required, how new internet technologies can be exploited and can benefit the promulgation, development, maintenance, and robustness of the standards. This paper is organized as follows. Section One introduces the concept of the virtual institutes, the expected benefits, and the motivation for our research and for research in this area. Section Two presents background material and a discussion of topics related to VIs, uman behavior and cultural modeling, and network-centric warfare. Section Three contains a discussion of the security challenges that face the virtual institute and the characteristics of the standards that must be employed. Section Four contains our proposal for documentation of the cybersecurity standards. Section Five contains the conclusion and suggestions for further work.
Programming by demonstration (PBD) is a technique for teaching the computer new behavior by demonstrating actions on concrete examples. This technique is intensely applied in systems for the creation of animated simulation, in order to make their interface more intuitive. This work proposes three improvements for animated simulation tools based on programming by demonstration: first person perspective for rules, use of inheritance and rules with negative conditions. The improvements suggested are implemented in a new system developed by the authors. The new system is compared with the most important similar tools, showing the benefits provided by the techniques proposed.
This paper reports the use of nuclear plant’s simulation for online dose rate monitoring and dose assessment for personnel, using virtual reality technology. The platform used for virtual simulation was adapted from a low cost game engine, taking advantage of all its image rendering capabilities, as well as the physics for movement and collision, and networking capabilities for multi-user interactive navigation. A real nuclear plant was virtually modeled and simulated, so that a number of users can navigate simultaneously in this virtual environment in first or third person view, each one receiving visual information about both the radiation dose rate in each actual position, and the radiation dose received. Currently, this research and development activity has been extended to consider also on-line measurements collected from radiation monitors installed in the real plant that feed the simulation platform with dose rate data, through a TCP/IP network. Results are shown and commented, and other improvements are discussed, as the execution of a more detailed dose rate mapping campaign.
This effort was part of the DARPA Active Templates program (2000-2004) to revolutionize mission planning, mission execution, and related command and control processes. Extensive use is made of previous research in generative planning and learning, case-based and mixed-initiative plan adaptation, real-time integration of action and execution, and multi-agent control and learning. Technology was developed to support users in creating and managing template- based plan, allowing them to anticipate multiple contingencies and dynamically re-plan based on real-time sensory information. The research is grouped into the following themes: (1) Allocation of communications spectrum frequencies, (2) Extraction of plan rationale and the learning of planning templates, (3) New abstraction techniques for reinforcement learning to improve the efficiency of automatic control algorithms, (4) Opponent modeling in dynamic multi-agent environment, (5) Multi-agent learning and limitations, and (6) Planning using symbolic model-based techniques. An extensive bibliography is included listing publications which describe the results of these research tasks in more detail.
Autonomous robot docking requires accurate path following and ultimately, accurate alignment with the target location, typically a docking station. In this paper we develop a novel, partially heuristic, technique that allows accurate docking for Ackerman-steered vehicles. The technique is validated via simulation. Multiple experiments were performed to better understand and analyze the heuristic element of the technique. The results underscore the impact of the vehicle's steering characteristics in docking precision and may prove valuable in attempting to remove the heuristic element.
RoboCup Challenge offers a set of challenges for intelligent agent researchers using a friendly competition in a dynamic, real-time, multi-agent domain: synthetic Soccer. While RoboCup in general envisions longer range challenges over the next few decades, RoboCup Challenge presents three specific challenges for the next two years: (i) learning of individual agents and teams; (ii) multi-agent team planning and plan-execution in service of teamwork; and (iii) opponent modeling. RoboCup Challenge provides a novel opportunity for researchers in planning and multi-agent arenas --- it not only supplies them a concrete domain to evalute their techniques, but also challenges them to evolve these techniques to face key constraints fundamental to this domain: real-time and teamwork.
This paper describes Gamebots, a multi-agent system infrastructure derived from an Internet-based multi-player video game called Unreal Tournament. The game allows characters to be controlled over client-server network connections by feeding sensory information to client players (humans and agents). Unlike other standard test-beds, the Gamebots domain allows both human players and agents, or bots, to play simultaneously; thus providing the opportunity to study human team behavior and to construct agents that play collaboratively with humans. The Gamebots system provides a built-in scripting language giving interested researchers the ability to create their own multi-agent tasks and environments for the simulation. Keywords Multiple Agent Systems, Agent Infrastructure, Multi-Agent Teamwork, Human-Agent Interaction, Virtual World Simulation, Video Game Bot. 1.
Unreal tournament: Action game of the year, 1999. GameSpot; www. gamespot.com/features
- J Gerstmann
Gerstmann, J. Unreal tournament: Action game of the year, 1999. GameSpot; www. gamespot.com/features/1999/p3_01a.html
Veloso (veloso@cs.cmu.edu) is an associate professor in the
- M Manuela
Manuela M. Veloso (veloso@cs.cmu.edu) is an associate professor
in the Computer Science Department at Carnegie Mellon University.
Master's Thesis, Centre for Virtual Environments
- T Evans
- Cluedo
Evans, T. Cluedo. Master's Thesis, Centre for
Virtual Environments, ISI, University of Salford,
2001.
Marshall (amarshal@isi.edu) is a staff member in the Information Sciences Institute at the University of Southern California
- N Andrew
Andrew N. Marshall (amarshal@isi.edu) is a staff member in
the Information Sciences Institute at the University of Southern
California.
Kaminka (galk@cs.cmu.edu) is a post doctoral fellow in the
- A Gal
Gal A. Kaminka (galk@cs.cmu.edu) is a post doctoral fellow in
the Computer Science Department at Carnegie Mellon University in
Pittsbugh, PA.
Unreal tournament: Action game of the year , 1999 . GameSpot ; www. gamespot.com/features/1999/p3_01a.html Gerstmann, J. Unreal tournament: Action game of the year
- J Gerstmann
- Gerstmann J.
Master's Thesis Centre for Virtual Environments ISI University of Salford
- T Evans
- Cluedo