1
Introduction

In response to a request from the Defense Modeling and Simulation Office, the National Academy of Sciences/National Research Council (NAS/NRC) is undertaking an 18-month study of the state of the art in human behavior representation as it applies to military simulations. Working within the context of the requirements established by military simulations, the panel is reviewing and assessing whether and how well the processes and effects of human behavior at the individual, unit, and command levels can be incorporated into these simulations.

A two-phase effort is under way. This report is the result of the first phase, which has been conducted over an eight-month period. As an introduction to its task, panel members attended a set of three workshops sponsored by the Defense Modeling and Simulation Office, at which leading military contractors described their efforts to model human behavior for a variety of military simulations. The panel has also had extended discussions with military modelers and, taking advantage of the expertise found in its membership, has begun an intensive exploration of the scientific domain of human behavior to identify areas in the literature that are most pertinent to the problems of military modeling. Based as it is on a relatively brief exposure to the problem area, the report provides a preliminary assessment of the needs of military users, suggests an approach to developing human behavior representations, and identifies some promising recent developments. The material presented here should be considered a statement of the panel's progress at the midpoint of its work. We attempt to summarize our present views concerning the relationship between human behavior representa-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 5
--> 1 Introduction In response to a request from the Defense Modeling and Simulation Office, the National Academy of Sciences/National Research Council (NAS/NRC) is undertaking an 18-month study of the state of the art in human behavior representation as it applies to military simulations. Working within the context of the requirements established by military simulations, the panel is reviewing and assessing whether and how well the processes and effects of human behavior at the individual, unit, and command levels can be incorporated into these simulations. A two-phase effort is under way. This report is the result of the first phase, which has been conducted over an eight-month period. As an introduction to its task, panel members attended a set of three workshops sponsored by the Defense Modeling and Simulation Office, at which leading military contractors described their efforts to model human behavior for a variety of military simulations. The panel has also had extended discussions with military modelers and, taking advantage of the expertise found in its membership, has begun an intensive exploration of the scientific domain of human behavior to identify areas in the literature that are most pertinent to the problems of military modeling. Based as it is on a relatively brief exposure to the problem area, the report provides a preliminary assessment of the needs of military users, suggests an approach to developing human behavior representations, and identifies some promising recent developments. The material presented here should be considered a statement of the panel's progress at the midpoint of its work. We attempt to summarize our present views concerning the relationship between human behavior representa-

OCR for page 5
--> tion and military simulations and to sketch out at least the beginning of an overall research strategy, to be refined and elaborated in the final report. In a second phase, the panel will conduct a more thorough review and analysis of the pressing issues in the current generation of battle management simulations and the theoretical and applied research of human behavior modeling as it applies to the military context at the individual and unit levels. The report prepared at the end of the second phase will recommend a research and development agenda designed to move the field forward in a systematic and integrated manner. This final report will offer guidance to researchers and practitioners who are developing military simulations as well as to those individuals who ate responsible for providing the research and development framework for future military simulation activities. Discussion of human behavioral representation among the experts working in this domain ranges much more broadly than is represented by the charge to this panel. The panel's focus is on the technology and knowledge needed to develop useful and usable human behavior models, ranging from the individual combatant to battalion and brigade-level units. The panel also discusses the analysis required as a prerequisite for model development, as well as verification and validation, because they are important to the generation and success of such models. The state of the an of computer technology, scenario generation mechanisms, and human interfaces to the models themselves are considered outside the panel's scope and are not discussed. On the basis of its work so far, the panel concludes that: (1) human behavior representation is essential to successful use of both war games and distributed interactive simulation; (2) current human behavior representations can be improved by incorporating knowledge of what is already known from the social and behavioral sciences, cognitive science, and human performance modeling; and (3) great additional progress can be expected through the funding of new research and through the application of existing research in areas that the panel is charged to identify in this interim report and in its final report. What is Human Behavior Representation? In general, the term model has different meanings in different communities. For some, a model is a physical replica or mock-up. For others, a model can be a verbal or analytical description or a block diagram with verbal labels. The term human behavior representation has been coined by the defense modeling and simulation community to refer to models of the human behavior or performance that needs to be represented in military simulations. In today's armed services, human behavior representations are being used for many different purposes. The prime beneficiaries of improved behavioral representations are the users of simulation, who can be categorized according to the specific application of the simu-

OCR for page 5
--> lations they use. Training simulation users are instructors and trainees who use simulations for instruction. Mission rehearsal simulation users are operational forces who use simulations to prepare for specific missions. Analysis simulation users employ their simulations to evaluate alternative weapon systems, doctrine, and tactics. Acquisition simulation users are those who must make acquisition decisions based on the anticipated performance of weapons systems. Command decision-aiding simulation users are commanders who wish to evaluate alternative courses of action before committing forces. Of course, the modelers who implement the models also stand to benefit from the availability of improved representations. Two primary critics will view the outputs of human behavior representation and judge whether they are credible or not. First, participants in real-time battlefield simulations will see the performance of individual soldiers and higher-level units in terms of the individual and unit behavior they exhibit, the execution of plans they formulate, and the battle outcomes that result. Second, players in non-real-time battlefield war games will observe only the resulting movements of troops and units, attrition results, and battle outcomes. The details of how individual humans actually accomplish tasks—that is, the psychological theory—is at this level irrelevant as long as the behavioral outcomes seem realistic. Similarly, many details about how groups accomplish tasks, such as group-level decision-making details, may be irrelevant. However, other details—such as, how communication works and the role of the command structure—may be important if the behavioral outcome is to be accurate. What is meant by realistic is open to interpretation and is discussed in more detail below. In this report we use the term human behavior representation to mean a computer-based model that mimics the behavior of either a single human or the collective action of a team of humans. It may be in the context of a self-contained computer simulation used to simulate a battle and run once or many times to produce outputs that reflect the battle outcomes, either individually or statistically. Or it may be a real-time simulation of the behavior of selected battlefield elements that can be viewed by real crews performing in other battlefield element simulators, such as squads of individual soldiers, ground vehicles, and aircraft, to allow the battle to be played out in the simulated world interactively. To review the state of the art in human performance modeling, with specific attention to potential military applications under the purview of the Defense Modeling and Simulation Office, is especially challenging because the ways in which the models will be used are substantially different from the goals and purposes of typical academic researchers studying and modeling human performance. Most academic researchers concerned with human performance are interested in the interplay between empirical data (experimental, field, or archival) and theory. They implement their theories in terms of executable models so that the detailed assumptions will be revealed and so that they can validate and evaluate the implications of the theory. Their theories are at the individual level and

OCR for page 5
--> are typically about specific human performance capacities and limitations, such as attention, decision making, and perceptual-motor performance. Rarely do these researchers articulate a comprehensive model of human performance that will behave in the aggregate like real humans. Nevertheless, that is the challenge presented by the requirements of military simulations. At the unit level, theories typically involve group performance and the way in which it is affected by the communication among group members, procedures, command and control structures, norms, and roles. Many of these theories can be articulated as computational models, which often illustrate the type of impact that an isolated change in procedures or structures may have. These models may generate realistic unit-level behavior; they are rarely accurate, however, at the individual level. The panel has been and will continue to be challenged by the need to focus on behavioral outcomes and to connect knowledge and theory of human behavior with realistic behavioral outcomes rather than becoming bogged down in details of theory. However, it is the panel's underlying concern that achieving the desired outcomes with both realism and generality requires the models to be based on the best psychological and sociological theory available. In fact, this is a limitation of the current modeling efforts the panel has reviewed. In the absence of such theory, the models are brittle in the sense that mild deviations from the conditions under which they were created produce unrealistic behavior and simplistic responses that do not correspond to the behavior of real individual soldiers or units. To avoid this brittleness and lack of correspondence between the model and real behavior, it is necessary to approximate the underlying psychological structure correctly. For example, in a simulation of the behavior of a flight of attacking helicopters, the helicopters were moving out to attack. One was designated the scout and moved ahead out of sight while the others hovered, waiting for a report. The scout was shot down. Having no further instructions, the others continued hovering until they ran out of fuel. The model could obviously be fixed to eliminate this specific bug in the program, by concatinating further if-then rules. However, what is really needed is a more general decision-making process for the lead pilot that can select among alternative courses of action when expected information does not become available as needed. Existing theory is the best means we have to define such situations. We also discuss models of learning as frameworks within which to build specific behavioral models. These models are often based not on an explicit theory, but on an architecture that is broader than the specific behavior to be simulated. However, operating within these frameworks may be helpful and important in minimizing brittleness. We discuss probabilistic and stochastic modeling frameworks that can help capture the unpredictability and variability of real behavior.

OCR for page 5
--> Setting Expectations in the User Community In discussions the panel has had with various representatives of the user community, it has become clear that there is wide variation in the expectations of what users believe is possible in terms of generating human behavior that is realistic, creative, adaptive, and consistent with military doctrine. We strongly suspect that what can be achieved in the near term is much more limited than what some potential users expect will be possible. One purpose of this project is to elaborate those aspects of performance that the panel determines are implementable now, those aspects that require significant translation of scientific theory and principles to develop into components of computer-based behavioral models, and those aspects of behavior for which the behavioral and social sciences provide inadequate knowledge for developing realistic models in the near future. The goal is to describe in general terms the theoretical and operating principles of models that are applicable to human behavior representation over varying time horizons, from immediately to three to five years; to describe the research necessary to translate current theory into future models; and to describe the areas in which models are needed, but for which theory must be further developed before the users can expect to have models that adequately meet their needs. These objectives will be addressed in the final report. A fundamental problem that faces the human behavior representation community is the assessment of which of the many modeling requirements will make a difference in the resultant quality of the models, depending on the intended use of the simulation. As the panel has deliberated, it has become clear that a panel of social and behavioral scientists, operating as a consultative group, cannot set these priorities without much more experience in dealing with the specifics of concern to the military community. We may be able to say what requirements will produce models that will more accurately characterize the somewhat abstract human processes of attention, memory, or decision making, but that is different from specifying the requirements that will produce models that appear to simulation users to be more like real individual combatants and military units. It is just not known which of the many improvements in human behavior representation will really make a difference in how a modeled combatant will be viewed, in terms such as meeting the expectations of the opposing force or minimizing the ability to ''game'' or learn the internal workings of the simulation. This is an aspect of the classic problem of simulation fidelity. Analysts would like to have high fidelity only where it matters, but for representing humans in combat simulations, no one has shown clearly just what aspects of fidelity matter.