Mainstream visual psychology presents a 'sense then infer' account of vision that is analogous to the 'sense then infer' processing that characterises the agent intention recognition literature. From ecological psychology comes Gibson's theory of visual perception that highlights the importance of the environment in explaining the nature of vision and recognition and claims that higher order structures are directly accessible. This theory can be used as the stepping-off point for an account of intention recognition and the means by which it might be modelled. Furthermore, the capacity for virtual environments to be designed 'agent friendly' provides yet another dimension of design freedom. When accompanied by an explicit model of perception the intention recognition problem can be cast as a software design problem. The resulting design patterns provide useful options for modelling intention recognition in intelligent agent systems. When constructing military simulations that require sophisticated cognitive models there are plenty of challenges to occupy the developer. Human factors experts, particularly cognitivists, endeavour to gain an understanding of actual human psychology behind the behaviours to be modelled; computer scientists attempt to develop computational technologies with the attributes necessary to model those cognitive functions. Ultimately engineers must draw these models and technologies together in a manner that results in simulation systems that meet the expressed requirements. In the world of military simulation requirements vary substantively. High fidelity, whilst seemingly always desirable often conflicts with performance, maintainability, complexity, and other attributes that combine to make the software less useful. Balancing these requirements is aided by the availability of several architectures that allow the engineer the freedom to tradeoff various attributes of the system. This thesis presents a series of design patterns that provides software architectures useful for implementing intention recognition. Each of these architectures has a basis (although sometimes this tenuous) in psychology and cognitive modelling, and each imposes requirements on the technology necessary for implementation. The relative merits of the patterns are presented as are fully worked examples of their application to flight simulation. APPROVED FOR PUBLIC RELEASE in DSTO-RR-0286IV DSTO-RR-02M Author Clint Heinze Air Operations DivisionClint Heinze is a cognitive scientist interested in the research spaces where software engineering overlaps AI. Primarily this involves attempts to discover ways of engineering representations of intelligence that have the properties commonly desired of quality software -that is they should be robust, reliable, validated, etc. He has a degree in Aerospace Engineering from EMIT and has recently completed his PhD in the Department of Computer Science and Software Engineering at the University of Melbourne. Since 1989 the Defence Science and Technology Organisation have em...

This paper reports experiences and outcomes of designing and developing an agent-based, autonomous mission control system for an unmanned aerial vehicle (UAV). Most UAVs are not truly autonomous or even unmanned but are more correctly termed 'uninhabited' or 'remotely piloted'. This paper explores two quite different approaches for adding autonomous control to an existing UAV. Both designs were implemented using an agent-based language. The first takes a fairly standard approach, adding a layer over the flight control system to control the mission. The second takes the human metaphor of agency more seriously and implements an autonomous controller based on a model of human decision making widely referenced in the military command and control literature. Implementing these two designs allowed a comparison of their relative strengths and weaknesses. Preliminary findings indicate both the feasibility and usefulness of a human cognitive modelling approach to providing autonomous UAV control but indicate a number of important considerations. This paper also reports on the successful first flight trials of the Codarra Avatar UAV under the mission control of the agent and discusses the future flight test program.

In Australia, the Scientists-in-Schools program partners professional scientists with teachers from K-12 schools to improve early engagement and educational outcomes in the sciences and mathematics.  An overview of the developing syllabus of a K-6 course resulting from the pairing of a senior AI researcher with teachers from a K-6 (primary) school is presented. Now entering its third year, the course introduces the basic concepts, vocabulary and history of science generally and AI specifically in a manner that emphasises student engagement and provides a challenging but age appropriate syllabus. Reflecting on the course at this time provides an action research basis for ongoing maturation of the syllabus, and the paper is presented in that light.

Mainstreaming and industrialising agent technologies requires suitable methodological and technological support for the various engineering activities associated with managing the complexity of any software system development. Despite its origins in object oriented software engineering the UML provides a rich and extensible set of modelling constructs that can be applied to agent oriented technologies. This paper provides details of extensions to the UML for the design of agents that are to be implemented in the JACK language. These extensions provide the capacity to model the behaviour of agents for the purposes of design and, though the extensions are language specijic, future generalisation and application to other agent languages can be supported as a industry-wide consensus about the nature of agency emerges over the next few years. This research builds on previously proposed extensions to the UML and moves a step closer to the goal of providing through-life engineering support to agent oriented systems development. This work is motivated by a pressing need to maintain, modi h , develop and deploy existing andfuture agent based simulations of military operations for the Australian Defence Force.

No abstract