Jeffrey Tweedale

A neural network-based multi-agent classifier system

In this paper, we propose a neural network (NN)-based multi-agent classifier system (MACS) using the trust, negotiation, and communication (TNC) reasoning model. The main contribution of this work is that a novel trust measurement method, based on the recognition and rejection rates, is proposed. Besides, an auctioning procedure, based on the sealed bid, first price method, is adapted for the negotiation phase. Two agent teams are formed; each consists of three NN learning agents. The first is a fuzzy min-max (FMM) NN agent team and the second is a fuzzy ARTMAP (FAM) NN agent team. Modifications to the FMM and FAM models are also proposed so that they can be used for trust measurement in the TNC model. To assess the effectiveness of the proposed model and the bond (based on trust), five benchmark data sets are tested. The results compare favorably with those from a number of classification methods published in the literature.

The Link between Agent Coordination and Cooperation

Agent coordination is the ability to manage the interdependencies of activities between agents while agent cooperation is the process used for an agent to voluntarily enter a relationship with another to achieve a system derived goal. We describe and show the concepts of Coordinative Cooperation and Cooperative Coordination using examples. These concepts demonstrate the ability for intelligent agents to distinguish between cooperation from coordination and vice-versa. Both concepts can be integrated into a process, using a cognitive cycle to explain the interaction between coordination and cooperation. Furthermore, this paper will discuss how the coordination/cooperation loop is initialised and can be affected by Coordinative and Cooperative events. We recommend suggestions on how these concepts can be designed and implemented in a multiagent system (MAS) and introduce AC3M, which is a prototype of this cognitive loop.

Fuzzy control loop in an autonomous landing system for Unmanned Air Vehicles

Humans are still responsible for operating complex equipment because they are able to respond during emergency situations. By removing the pilot from an aircraft, Unmanned Air Vehicles (UAVs) are vulnerable to unpredictable activity within their environment. The landing process is the most stressful phase of flight for pilots and as a consequence, a significant number of UAV are being destroyed or unnecessarily damaged. An automated landing system similar to that for civil aviation would not only improve safety, but it would reduce the cognitive loading of operators and logistic costs. A small team will integrate a segmented Fuzzy Logic Controller (FLC) with a low cost sensor (The XBox 360 Kinect) to create an autonomous landing system for UAV platforms. The knowledge discovered while processing and analysing the data captured during the course of the project will be used to evolve further capability into the controller.

Future Directions: Building a Decision Making Framework Using Agent Teams

1 School of Electrical and Information Engineering, Knowledge Based Intelligent Engineering Systems Centre, University of South Australia, Mawson Lakes, SA 5095, Australia 2 Air Operations Division, Defence Science and Technology Organisation, Edinburgh SA 5111, Australia 3 The Sellinger School of Business and Management, Loyola College in Maryland, Baltimore, MD 21210 USA 4 School of Electrical and Information Engineering, University of South Australia, Adelaide, South Australia SA 5095, Australia

Trust in multi-agent systems

Research in Multi-Agent Systems has revealed that Agents must enter into a relationship voluntarily in order to collaborate, otherwise that collaborative efforts may fail [1,2]. When examining this problem, trust becomes the focus in promoting the ability to collaborate, however trust itself is defined from several perspectives. Trust between agents within Multi-Agent System may be analogous to the trust that is required between humans. A Trust, Negotiation, Communication model currently being developed, is based around trust and may be used as a basis for future research and the ongoing development of Multi-Agent System (MAS). This paper is focused on discussing how the architecture of an agent could be designed to provide it the ability to foster trust between agents and therefore to dynamically organise within a team environment or across distributed systems to enhance individual abilities. The Trust, Negotiation, Communication (TNC) model is a proposed building block that provides an agent with the mechanisms to develop a formal trust network both through cooperation1 or confederated or collaborative associations 2 . The model is conceptual, therefore discussion is limited to the basic framework.

A future framework for interfacing BDI agents in a real-time teaming environment

Research and applications in human-machine teaming continue to evolve the role of the human from immediate (manual) operator into supervisory and televisory controller. In the supervisory control role, the human operator will be functionally removed from the system under control and in the televisory role, the human operator will be physically removed. Although unmanned systems and vehicles have become a technical reality that drives this change, they will not eliminate the importance of the human operator as the commanding and controlling element in-the-loop. This paper will argue that existing automation concepts remain equally valid with an even greater emphasis on the need for a human-centered automation approach. Intelligent agent technology has become mature and attractive enough to implement the automated components of the human-machine team. Agents that implement the Beliefs-Desire-Intention syntax will be discussed as being of particular interest for human-machine teaming applications. This paper proposes a theoretical framework for teaming human and intelligent agents. The teaming framework will be demonstrated in a real-time simulation environment using the commercial game called Unreal Tournament and its existing GameBot extension. The intelligent agents will be implemented based on the Belief-Desire-Intention (BDI) syntax and using JACK, a commercial BDI Agent development language. The requirements for follow-on research, such as human-agent teaming, human-agent coordination and agent learning will be highlighted.

An overview of agent coordination and cooperation

By their very nature, intelligent agents possess four important social abilities. These include the ability to communicate, cooperate, collaborate and the need to be coordinated. This paper presents an overview of two of these social abilities, that of being coordination and cooperation. The discussion develops the theory of each and derives the current definitions. The definitions will then be linked into a single multi-agent system (MAS) model, Agent Coordination and Cooperation Cycle Model. This shows a cognitive loop that replicates the link between coordination and cooperation in systems such as organizations, management and biological systems. This paper will also present the advantages, consequences and challenges associated with the implementation of Agent Coordination and Cooperation Cognitive Model (AC3M) within intelligent multi-agent systems.

Intelligent Agents and Situation Awareness

Advances in automation and artificial intelligence, and especially in the area of intelligent (machine) agents, have enabled the formation of rather unique teams with human and machine members. This paper describes initial research into intelligent agents using a 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. Intelligent agents using the BDI-reasoning model can be used to provide a situation awareness capability to a human-machine team dealing with a military hostile environment. The implementation described in this paper is using JACK agents and “Unreal Tournament” (UT). JACK is an intelligent agent platform while UT is a fast paced interactive game within a 3D-graphical environment. Each game is scenario based and displays the actions of a number of opponents engaged in adversarial roles. Opponents can be humans or agents interconnected via the UT games server. To support research goals, JACK extends the Bot-class of UT and its agents apply the BDI architecture to build situational awareness. This paper provides the background for the use of intelligent agents and their cognitive potential. The research is described in terms of the operational environment and the corresponding implementation that is suitable for intelligent agents to exhibit BDI behaviour. The JACK application will be described and specific requirements will be addressed to implement learning in intelligent agents. Other cognitive agent behaviour such as communication and teaming are aims of this research.

A neural network-based multi-agent classifier system with a Bayesian formalism for trust measurement

In this paper, a neural network (NN)-based multi-agent classifier system (MACS) utilising the trust-negotiation-communication (TNC) reasoning model is proposed. A novel trust measurement method, based on the combination of Bayesian belief functions, is incorporated into the TNC model. The Fuzzy Min-Max (FMM) NN is used as learning agents in the MACS, and useful modifications of FMM are proposed so that it can be adopted for trust measurement. Besides, an auctioning procedure, based on the sealed bid method, is applied for the negotiation phase of the TNC model. Two benchmark data sets are used to evaluate the effectiveness of the proposed MACS. The results obtained compare favourably with those from a number of machine learning methods. The applicability of the proposed MACS to two industrial sensor data fusion and classification tasks is also demonstrated, with the implications analysed and discussed.

Intelligent Decision Support System in Defense Maintenance Methodologies

Integrated logistics support (ILS) systems are becoming more sophisticated as the defense operations demand increasing levels of service from contractors and maintainers. Intelligent decision support system (IDSS) are required to provide adaptive automated responses for provisioning and maintenance of an increasing number of defense platforms that are now being supported either by defense organizations or by long term maintenance contracts from within the private sector. Traditional methods of repair are being upgraded to include data mining and intelligent agents to help stimulate intelligent decision support systems in creating a more efficient and reliable maintenance environment for defense. A concept demonstration, called the automated test equipment multi-agent system (ATEMAS), is being developed under a collaborative program between Defense Science and Technology Organization (DSTO), the University of South Australia and Raytheon Australia. It is envisaged that when developed, this system will be able to provide cognitive intelligence to support better reliability predictions and thus making possible the much needed proactive obsolescence management for avionics parts