SBIR-STTR Award

Using distributed interactive simulation (DIS) to conduct effective training, analyses, or mission support in a tactical nuclear weapons effects (NWE) environment requires realistic computer generated forces (CGF) that are sensitive to such effects. This level of sensitivity requires CGF to apply valid and reliable speed-accuracy measures of environmental effects on human performance within the DIS environment. a major deficiency in DIS and CGF is that they do not currently simulate the interaction between response time and accuracy that is so characteristic of the performance of humans under stress. Our proposed effort will address that void to enable a more realistic representation of this dynamic dimension of human performance. To fully characterize human behavior in quantitative terms, we need better expressions of task performance accuracy. Unlike response time, operational definitions of accuracy vary for each type of task (e.g., cognitive, psychomotor, and physical). In addition, response time and accuracy measures are not independent. Often the human op0erator must decide which aspect of the task is more important and then trade speed of response for accuracy as required. To develop such a speed-accuracy model (SAM) PSR will analyze human response to NWE based on our in-depth studies of human behavior. We will also investigate the extensive literature on speed-accuracy interaction and compare the utility of alternate modeling approaches to specific human response requirements. Finally, PSR will plan a feasibility demonstration of our SAM algorithms for use within DIS. In air traffic control simulations, SAMs of pilot, crew, and controller responses under high stress would provide a more credible environment for training controllers. In nuclear power plants or other process control facilities, SAMs of control room personnel performance during emergencies could be used to train managers or analyze incidents for lessons learned. In the entertainment industry, opponents would be more challenging if they exhibited a wider variety of behaviors and stress-induced performance decrements.

Realistic computer generated forces (CGF) such as Modular Semi-Automated Forces (ModSAF) requires a realistic representation of human behavior which must reflect the speed versus accuracy (SVA) trade-off in task performance. A major deficiency in today's CGFs is that they do not currently simulate the trade-offs between response time and accuracy that is so characteristic of the performance of humans operating in a stressful environment. The purpose of this proposed Phase II PSR effort is to address the deficiency and fill this recognized void in our current approach to human performance modeling. The Phase I effort provided a foundation for establishing higher credibility ad validity in the behavior of synthetic forces operating within larger scale simulations. An in-depth understanding of the SVA trade-off was gained through an exhaustive literature search represented in terms of mathematical equations. In Phase II, the modeling equations will be formalized including the development of appropriate parameter values from empirical data, testing, evaluation, verification, validation, and accreditation (VV&A) Speed-accuracy models (SAMs) of "first responders" to a chemical or biological agent release in civilian sector would support more valid threat analysis and preparedness training for emergency response actions. SAMs of pilot, crew, and controller behavior under high stress would provide a more credible environment for selecting, evaluating, and training air traffic controllers. Similar benefits apply for personnel operating nuclear power plants or other process control facilities.

Keywords:
Human Performance, Accuracy Measures, Modsaf, Response Time, Distributed Interactive Simulation, MOD