Applications

Realistic Virtual Actors

Tue, 19 Jan 2010 12:14:54 +1100

The current generation of video games provides a highly attractive immersive gaming experience, combining high quality animation and visual effects. Some games rely upon “virtual actors” or avatars to populate the games, whether they are the enemy or friendly characters. The virtual actors’ behaviours are developed for each scene, and all their decisions and consequent actions are implemented for each case as Finite State Machines (FSMs).

However, it is very time consuming to develop the actors and their behaviours, which are often a limited repertoire of scripted actions. Using CoJACK™ to control the virtual actors greatly improves the range and credibility of their behaviour, while reducing the programming effort involved.

Military forces around the world have historically used bespoke computer simulation systems, or Synthetic Environments/Battlespaces, to study the effectiveness of their tactics and equipment when engaged with potential opposition forces. Recently this situation has changed dramatically. The availability of low-cost, very high performance desktop PCs has allowed software developers to produce a new generation of photo-realistic military training systems. These provide a compelling immersive visual experience and have been recognised by the military as offering a significant improvement in affordability for training systems, particularly for land forces.

Human Behaviour Representation

Thu, 08 Nov 2007 18:54:39 +1100

In most current synthetic environments, the individual entities in the simulation will execute the same task in the same way, ignoring differences between individuals, and even the variability of a given individual over time. In the real world, this is not the case. The choice of strategies and the ordering of sub-strategies will vary across individuals and will vary for a given individual across time. When such variance is not included in a model, it makes adversaries, allies, and neutral personnel too predictable because they will always do the same thing at the same time in the same way. Recent thinking recognizes the inherent variability of humans (i.e. differences in cognition and physiology from one individual to the next), and the necessity of modelling this to improve the realism of synthetic forces.

The Improved Human Behaviour Representation (IHBR) project (UK Ministry of Defence) addresses this shortcoming. Using CoJACK™, realistic human variation is generated in a psychologically-grounded and therefore principled manner. This is a major step forward relative to the ad hoc approaches of the past. CoJACK predicts variation in decision-making time as well as errors that can occur due to moderating factors (e.g. fatigue). Situation awareness is also explicitly represented and logged, allowing analysis of the extent to which unexpected behaviour resulted from deficiencies in situation awareness.

Oil and Gas

Thu, 08 Nov 2007 13:46:31 +1100

Norwegian-based Statoil, one of the world’s largest suppliers of crude oil and natural gas, has developed software to support oil trading and operations management, using JACK.

Intelligent agents are applied to help solve optimisation, planning and process control issues in Statoil's trading and operations areas. Among the expected benefits in the trading area is the timely provision of the most up-to-date information, enabling the company to rapidly respond to changes in the market. This speed of operation allows Statoil to improve effectiveness and continually modify ongoing business tactics, giving the company ample opportunity to further maximise revenues, while reducing the number of decisions made on out-dated information.

Statoil chose JACK as its development platform because of AOS’s field-proven track record in knowledge engineering applications, its extensive training and mentoring services, and the availability of valuable support from Cambridge Consultants.

"To prosper in many of today’s global markets companies need to gather and act on only the latest information. By integrating agent technology into their system, Statoil can focus on running its business, safe in the knowledge that these programs are gathering business-critical data all the time", says Cambridge Consultants' Grant Garner.

Rules of Engagement

Thu, 08 Nov 2007 12:39:11 +1100

The ROE3 application (Rules of Engagement Evaluation Environment) adds sophisticated modelling of ROE (Rules of Engagement) to simulation environments. In the past, the modeling of ROE in a simulation context has been fairly rudimentary (e.g., setting fire permissions and target priorities in the OneSAF Testbed Baseline).

A key problem in modelling how military personnel behave during a conflict is to understand how ROE are interpreted with respect to the tactical environment and mission goals to be achieved. Until ROE3, the following factors were not modelled:

ROE,
how a person selects a tactic within the scope of several ROE being applied simultaneously, and
how this selection process is affected by the type of operation, psychological aspects of the person, and moderating influences experienced by that person (such as stress or fatigue).

Whether during drafting, training or conflict, ROE present a variety of challenges and tradeoffs. For example, ROE designed to limit escalation can put friendly forces at risk, particularly where the enemy has no formal ROE. Evaluating draft ROE is a major challenge in the absence of a conflict arena.

ROE3 offers a number of significant benefits:

Candidate ROE can be systematically evaluated for effectiveness and unforeseen implications.
ROE3 encourages ROE formalisation standards, thus reducing the cost and time taken to evaluate changes to ROE.
ROE3 includes modeling human variability factors like stress and fatigue, thereby supporting the evaluation of how these moderators can affect ROE application in a wide variety of circumstances.

This application is described in more detail here: BRIMS 2007 Best Paper

Intelligent PHM

Fri, 02 Nov 2007 22:20:18 +1100

The Intelligent Prognostic Health Manager (iPHM), developed by AOS, is an autonomous system that takes prognostics to the next level of capability.

It combines existing fault diagnosis techniques with Estimated Time to Failure under various conditions and provides recommended courses of action - for example to a military pilot, ship's engineer or an onboard Mission Management System.

Utilising iPHM, autonomous systems can intelligently react and accommodate failures, allowing them to maximise their remaining capability.

iPHM can be applied to various systems on manned or autonomous air, ground, sea or underwater platforms, including Power Management Systems, Mission Management Systems, Sensor Systems and Communications Systems.

iPHM can also be utilised in commercial applications such as oil platforms, mining plants, merchant ships and on civil aircraft.