Architectures, Human behaviour modelling, Intelligent Agents, Interaction between Live and Simulation, Standardization, Training
Human behaviour modelling (HBM) includes the quantitative representation of performance, decision making and behaviour of individuals and small groups. It is an emerging technology with both wide a range of applications and the commensurate challenges to put these models reliably into practice as indicated in NATO Symposia (eg HFM-202) and Specialist Team meetings (eg MSG-107).
A particular area of interest is the seamless interaction of Live players with realistically simulated human characters. This capability has broad applications, ranging from training for urban operations, to red force representation in tactical Air-to- Air training. The required level of fidelity in such models may vary considerably across uses. However, they are all intended to represent important characteristics of human cognition and performance.
There is a need for standards for operational model architectures wherein humans and virtual humans can work together. These virtual humans should adequately model behaviour based on general human aspects (eg cognition, emotion, physiology) as well as on cultural background and role in society. The behaviour should be validated and fit-for-purpose to meet the requirements for military training. There is also a need for support of more natural interaction between human trainees and simulated characters (e.g. gestures and speech rather than keyboard and mouse).
Significant research is ongoing in this field (eg HFM-128). However, the NATO RTO, NATO armaments groups and the NATO military modelling, analysis and simulation communities would be well served by a consensus Reference Architecture for HBM. This would provide a common framework for developing HBMs for training, facilitating model reuse and information exchange, and ultimately savings in time and resources for development of particular training applications.
The challenge is the integration of science based models, that describe (only) part of human behaviour, into complex military training settings. The scope and limitations of currently developed models and architectures need to be investigated and the best approach for a consolidated Reference Architecture must be decided. Exploration of the state-of-the-art in human behaviour modelling, architectures and implementations requires a cross-disciplinary approach that should involve NATO experts from HFM (Human view), NMSG (architectures and standards), academia and industry.
• Analysis of relevant training applications
• Conceptual Modelling of individuals
- Social and cultural influences on behaviour
- Cognition, including decision making, error and planning, emotions
• Investigation of (Sub)model architectures and hierarchies
• Development of Reference Architecture capturing model integration and model interface standards
• Training provision validation and implementation
• Guidelines for tailoring the Reference Architecture to specific applications and implementations