PREFACE

Properly preparing its armed forces for future conflicts is one of the most difficult challenges facing any nation. This is particularly true during a time (such as the present), when entire geopolitical, cultural, economic, and technological landscapes are undergoing significant rapid change.

Our national command authorities and defense forces are faced with dramatic shifts from a bipolar world to an amorphous context, where the threats to national security and the potential fields of combat are no longer clearly predictable, Adding to the pressures on our defense establishment are realities of the economics of the post-Cold War world. The total impact of these changes is staggering. Our well defined force structure, forward bases and other warfighting infrastructure are being adjusted to meet the new challenges, yet we have reduced resources to respond to the evolving paradigm. The command, control, communications, computer, and intelligence (C41) infrastructure that was built to fight a predictable adversary in predefined areas was not designed to respond to the uncertain threats of the future. Coping with these national security challenges requires the effective exploitation of emerging technical changes based on re-engineering functional processes. This is particularly true in the area of information systems in the degree to which they can enhance the effectiveness and responsiveness of the command and control process. Emerging technology offers an opportunity not only to cope with current challenges but to enhance the overall posture and capabilities of the armed forces. Our challenge is to choose wisely from the vast array of emerging technologies and to adapt our concepts, organizations, and procedures to take advantage of the operational capability enhancements they enable.

Military strategy and tactic has often failed to adapt in a timely manner to technological advances of the day. Despite the advent of the rifled barrel, both armies of the Civil War suffered horrible casualties as they continued to fight Napoleonic en masse formations appropriate for delivery of inaccurate volleys of musket fire. The French were defeated handily in the opening days of World War II, as they continued to utilize a trench warfare strategy behind their Maginot Line, ignoring that mechanization had delivered an enemy fighting a highly maneuverable blitzkrieg-type battle. During the past two decades, a revolution in technology has completely redefined the manner in which information is collected, processed, and disseminated; the processing power of today's desktop personal computer was unthinkable at the beginning of this decade. Today's average consumers wear more computing power on their wrists than existed in the entire world before 1961. But, despite the exponential improvements in information and telecommunication systems that we experience today, military strategy and tactic is adapting only incrementally. If our nation is to capitalize on technology advances, expanding at exponential rates, and not adapt strategy, tactic and operational techniques and procedure in consonance with the improved technology, we will miss a rare opportunity to respond in a positive and effective way to the changes of the evolving paradigm. Therefore, we must strive to identify how to exploit and leverage (what some call) the "information technology revolution" by applying it as an effective part of our defense capability.

Change is difficult for all organisms. This is particularly true for a well organized and trained military force, where tradition, discipline, and doctrine are the very foundation and framework upon which readiness and proficiency are built. Nonetheless, change is essential for leadership in a changing world. Success will be measured in terms of how effectively we can select and implement changes to meet the needs of the day and how well we can adapt our doctrine, organization, and procedures to exploit opportunities for change.

We can apply technology as a simple "delta" improvement to currently accepted functional processes, or we can implement processes that exploit and leverage technology to radically improve defense capabilities

We must first understand what can be changed and where the greatest opportunities exist for leveraging defense capabilities through technological change and corresponding change in doctrine, equipment, facilities, organization, concepts, tactics, and related training and education. Next we must identify and implement the means to effect the changes and to keep them synchronized to obtain the full benefits. LTGEN Van Riper, USMC (Ret) has frequently contrasted the British use of armor in World War I and the German use of armor in World War II as an example. The British introduced tanks, but they used them as individual piecemeal adjuncts to the infantry force, principally to counter the machine gun. The Germans, on the other hand, introduced a radically new concept for warfare by combining the tank, the radio, and the aircraft to establish the "lightning fast" and massive power of the Blitzkrieg. The Germans exemplified the power that technology can bring to bear if properly matched by adaptation of doctrine, organization, and tactics. There were similar contrasts in the way that the United States Army and the German Army used tactical air power in the early stages of World War II. The catastrophe for U.S. forces at the Kazerine Pass caused us to review our doctrine, organization, and procedures for using air power and to adopt the German model for commanding and controlling air assets.

2.1 Technical Change

Applied science has enabled major advancements in three significant warfare functions (among others) over the past decade: sensing, command and control, and guidance of weapons. The computer chip has been the technological centerpiece for the increased capabilities in each of these. There is every indication that improvements will continue to be made in these three areas far into the future.

Sensors now allow commanders to "see" the enemy in near real time under conditions and at ranges never before possible. The most significant new capabilities are provided by those systems which sense ground targets. The Joint Surveillance Target Attack Radar System (JSTARS) is an example of such a sensor. Others less well known (generally due to the classified details of their capabilities) are carried by other manned and unmanned aircraft. The advantages these provide to targeting have been amply demonstrated in Iraq and recent operations in Bosnia. Evolutionary improvements are also being made to systems which sense targets in the air and under the water. However, probably the most important advancements are being made in processing and control systems that enable massive quantities of target data to be collated, correlated, fused and deliverd to operational forces as a common operational picture. In other words, the most significant advances are associated with the processes that provide inputs to the command and control functions. These inputs are more plentiful, more comprehensive, and more detailed than before and offer tremendous potential for operational commanders and weaponeers to cope with uncertain environments.

The value of precision guided weapons has become more evident over the past few years as enabling technologies have emerged as mature and reliable elements of modem military forces. These smart weapons are classic indicators of technology-motivated revolution in military thought and practice, They provide a significantly higher assurance that a target will be destroyed. Equally important, they reduce the risk to friendly personnel as well as the possibility of collateral damage. Though cost constrains the development of large numbers of precision weapons there are many more on the horizon. When they are added to our arsenals, they will further increase the ability of friendly forces to fire or launch from standoff positions in all weather conditions, day or night, with very high probabilities of destroying the target on the first attempt. However, the effective use of these assets depends on the ability to acquire, process, and act on information to target them and to control them from "sensor to shooter." This leads once more to the function of command and control.

Command and control has always been the most difficult of battlefield functions to perform effectively because of the inherent chaos and confusion of war. Nonetheless, since it is central to every other activity of war it has been (and remains) a principal concern of commanders. Therefore, it is not surprising that immense hope is placed in the potential of improved communications and advanced information technology to enhance command and control to enable dominance of future national security operations. This hope is expressed by the Joint Staff's Joint Vision 2010, and in such concepts and programs as "information warfare," "cyberwar," and "digitization of the battlefield." Command and control enhancements will benefit from the improvements in sensing and processing, since the input information will improve quality, quantity, and timeliness. The increased availability of satellite communications, the introduction of fiber-optic cable, and the wider use of advanced computers will undoubtedly allow for increased amounts of information to be moved rapidly around the battlefield and will help integrate command, both vertically and horizontally, including across Service and national seams. These will also provide increasing leverage to future operational forces through the expanded ability to control and coordinate firepower from organic and nonorganic assets, such as the precision strike systems.

Nevertheless, one of the more difficult tasks (and a major focus for this paper) is providing capability to manage, integrate, present, and use available information in a way that helps the commander make better decisions regarding the use of forces and weapons made available to US joint and international commands.

2.2 Functional Change

Means or "tools" of change to defense capabilities can be grouped into five categories: doctrine, organization, training and education, equipment, and facilities. These are interdependent and need to be adapted in an integrated manner to achieve change. As a general rule, modifying or developing new doctrine is the least costly means to bring about change while equipment is the most expensive, with the other means falling between. Often, however, the inclination is to pursue equipment solutions when problems or deficiencies are detected. Not only is this unwise fiscally, it ignores the fact that in the past changes in doctrine have generally reaped the largest dividends.

Our nation must maintain a clear focus and emphasis on the integrated evolution of all five categories. New technology needs to be implemented within the force as part of the restructuring of the way that the force is organized and operates. In industry, this is called "reengineering the enterprise." In the military world, the process amounts to much the same: a revolutionary change in the way we do business, built upon and supported by the revolutionary changes in the available tools. Affordability and effectiveness of revolutionary change directly relates to the integrated nature of its implementation. The next task is to identify the combat functions (and the manner in which they can be changed consistently) which will have the largest impact on defense capabilities at affordable costs.

2.3 Enhancing Warfighting Capabilities

Certainly changes can be applied to improve any or all of the four basic functions performed by combat units; maneuver, fires, logistics, and command.

Maneuver refers to the movement of combat formations over land and water, or in the air. The purpose of maneuver is to place the friendly force in a position of advantage relative to the enemy. The forms of offensive maneuver are limited to envelopment, turning movement, infiltration, penetration, and frontal attack. The two forms of defensive maneuver are mobile and area defense.

As LTGEN Van Riper, USMC (Ret) has emphasized in many lectures to senior defense officials, it is difficult to imagine additional forms of maneuver, when one considers the geometry of either a two or three dimensional battlefield. A unit could perhaps combine forms in unique ways, employ deception, and so forth, but other than moving somewhat faster and further its options are circumscribed by the rules of geometry and the laws of physics.

Mobility, the ability of units to move, impinges directly on maneuver. Once again, betterment of mobility seems to be constrained by geometry and the laws of physics as well as defacto limits imposed by cost-gai tradeoffs, General rules of thumb exist for maximum practical speeds for land, water, and air vehicles, given the constraints imposed by structures and human factors and by the power plant implications. We should not seek major cost effective enhancements to the force by some revolutionary improvement in maneuver. We can expect incremental improvements as well as synergistic improvements achieved by better coordination of maneuver capabilities with the other three basic functions.

Fires refer to the employment of weapons systems. In terms of ballistics, fires are either direct or indirect. The fires of a weapon are measured by volume (numbers of rounds or weight of explosives) and rate of fire. Weapons fire can also be measured by range and accuracy. Improvements in each of these measurements is theoretically possible, especially in accuracy, and considerable work is ongoing to ensure these improvements are achieved Except for accuracy, most gains made will most likely be incremental until electromagnetic or energy directed weapons can be fielded. One exception is the emerging capability to employ precision-guided weapons and long range stand-off weapons. This class of weapons will have significant impact but mainly for a limited class of targets: those that are of sufficient value to be attacked and are small enough in number to be engaged to an operationally significant degree.

Logistics consists of supply, equipment maintenance, and transportation, among other capabilities. Because resources will always be limited, effective logistics is also efficient logistics; both are realized by proper management. Strategic and tactical mobility, and visibility of assets affect logistics greatly. These two factors offer considerable room for improvement of battlefield logistics. Planned increases in strategic sealift and airlift, replacement of truck fleets, and improved automation of supply and spare parts systems will enhance logistics. Innovations in this functional area promise to reap large rewards in combat capabilities. New technologies for managing inventories and information and for controlling logistics movement may enable a new concept for "just enough, just in time" logistical support. This concept is attractive in its economic and logistical implications, but it is equally disturbing in its potential for catastrophic consequences if it should fail.

The command and control function permits a commander to direct the actions of his unit. Historically, command and control has been the most difficult of the combat functions to perform because of the inherent "fog" and "friction" of war. Effective command and control requires the right combination of personnel who, equipped with adequate communications and support equipment, judiciously apply proper doctrine to plan and execute operations. Few military forces have attained great proficiency in command and control. There are two reasons. First, the complexity of the task is often overwhelming. Second, insufficient attention has been paid to the fundamental conceptual aspects of the function, hindering efforts at improvement.

Historians, researchers, and observers of the military are nearly unanimous in their conviction that potential improvements to command and control can revolutionize military capabilities, Improvements in command and control, enabled by modern information technologies, can have spin-off effects in the other functional areas, Maneuver will be able to be applied more effectively to meet the situation. Fires will be able to be directed more effectively, and the coordination of-maneuver and fires will be enhanced. Logistics will also be better coordinated with real time needs and with projected needs as the commanders ability to assess the current and future situation improves and as his or her decision making improves. Command and control is perhaps the key function where leverage can be obtained to improve warfighting capability in any revolutionary sense at reasonable (or even reduced) costs.

A review of the four functions of combat reveals that new equipment can be expected to make only marginal improvements to maneuver and fires. By contrast, improved and added equipment can make an important contribution to bettering logistics. Such equipment is being sought in current budgets. Marked improvements to maneuver, fires, and logistics will depend heavily on enhancements to doctrine, organization, facilities, training and education.

Significant command and control capability improvements can be obtained from technological advances, but even more so from doctrinal and organizational changes. Successful integration across all means of change (doctrine, organization, training, education, equipment and facilities) would cause command and control to truly set the pace of a "REVOLUTION" in military affairs and national defense capabilities. The following sections outline how this can be accomplished.

Command and control consists of two principal elements: decision making, and execution and monitoring. The remainder of this paper will address the decision making process and focus on knowledge required by the decision maker. This is not intended to imply that execution is not equally important but only to bring attention to an area that is both interesting and important in terms of the opportunities for revolutionary improvements.

3.1 Command and Control Decision Making

According to Martin Van Creveld, in Command in War, the unifying concept underlying command and control decision making is the need to cope with uncertainty. Uncertainty stems from the fog and friction which prevents a commander from seeing and understanding the battlefield as it actually exists. The only way to reduce that uncertainty is to acquire accurate information for the commander that provides additional knowledge regarding the situation being experienced. The commander will then have an increased understanding of the battle situation. Acquiring and processing information takes time and, as the situation changes, the value of the time-late information tends to decrease. Also, there is the likelihood that information will be lost, corrupted, or deliberately distorted as it is acquired, processed, disseminated, and integrated.

Command and control includes within it the art of making decisions from sparse or conflicting information and acting upon those decisions in a way that best represents the commander's estimate of the situation and the alternatives. This has been called a "coup d'oeil".' To capitalize on and leverage this capability, C41 system designers have two alternative approaches from which to choose. They can attack the information availability problem head-on and try to capture and process as much information as possible, or they can reconcile themselves to a construct in which the information will remain incomplete and where the system and the commander are oriented toward coping with uncertainty.

LTGEN Van Riper, USMC (Ret) has made a "revolutionary" proposal here too; that the most appropriate way to proceed with C41 systems planning and design is to work from the premise of minimal information and to provide support that helps the commander make effective and timely decisions "in spite of uncertainty." This approach is also reflected in work done by Major John Schmitt, USMCR, who has built upon research by Dr. Gary Klein on real world decision making. Dr. Klein has demonstrated that most decisions in real life are made intuitively, not in a rigorous analytic fashion. Maj. Schmitt has suggested that the correct command and control (C2) paradigm would correspond to this type of behavior and would be organized to support an intuitive decision maker. The underlying message in the writings of these authors (listed in the attached bibliography) is that C2 system need to present information or knowledge to the decision maker in the form of easily understood "patterns" that map to the commander's intuitive thought processes. It is interesting to note that the "intuitive decision maker" falls into the Meyers-Briggs personality category known as the "field marshal personality". This may be a coincidence or a very perceptive statement of fact.

3.2 Knowledge and Experience - The Two Pillars of Decision-Making

As noted above, the commander will always have to cope with uncertainty and with the stress of warfare when making decisions, which will be based largely on the commander's judgment derived from experience. This influences intuitive behavioral responses. He or she will certainly use all available information and will attempt to interpret it and evaluate the potential implications of alternative courses of action. Nonetheless, when the decision is actually made, it will probably be made from a heuristic intuitive basis rather than a strictly analytical one. In other words, the commander will have to rely on more than complex computer algorithms to make the decision. Knowledge and experience will drive the intuitive decision process, The more competent the commander and the more complete the knowledge, the better will be the decision,

Figure 1 illustrates the two pillars of decision-making. On the one side, information is processed to knowledge and understanding. On the other, experience is expressed as judgment. The sum of the two is what produces the decision.

Thus, C41 systems that support the commander must be structured to provide information in the form of concise knowledge capable of being rapidly assimilated and understood. The commander must be able to quickly recognize the picture and query the system for underlying facts or estimates that intuition and training identify as being vital to the understanding. Wherever possible, the systems should also provide automated aids for filtering and correlating knowledge, and for making rapid projections and estimates. The commander will use these capabilities as tools to leverage his judgment in making decisions.

3.3 Coping With Uncertainty

As noted above the most powerful tools for making decisions without certainty are those that provide the types of knowledge that match the decision maker's intuitive thought patterns. The decision maker must have sufficient experience and good judgment to make intuitive decisions. Finally, the decision maker must have sufficient confidence in his or her own judgment and in the validity of the information that is presented. To achieve these objectives, C4I systems must be built in a way that allows users to "customize" their interactions with them. They must be able to organize and tailor the displays to suit personal preferences and to make queries and interactions that satisfy their needs for developing understanding and confidence,

C41 systems also must support experiential learning so that users can develop their style of interaction with the system prior to the actual use in warfare. When real-world use is at hand, the commanders and staffs must already be proficient and confident with using the tools available to them. Hence, there is need not only for high capability systems that are tailored to the commanders' decision making processes; there is also need for embedded training and practice capability that will allow them to become confident experts in using the systems under realistic conditions of uncertainty and stress.

The following section on the decision making process addresses the fundamental issue of moving from information to knowledge, then to understanding, and then to a reasoned decision from experience-based judgment.

A commander must interpret a situation in terms of status, threats, capabilities, and objectives and choose and execute a course of action that best meets the objectives. The commander chooses among a number of courses of action based on a best estimate of the likely results. The basic resource used to support decisions is information. In principle, the more information available, the better the estimate assuming of course that the information is in a form capable of being rapidly assimilated and understood in context of knowledge relating to its intended use.

4.1 A Simple Model for Decision Making

A simplified model for decision-making is shown in Figure 2 and will be as a reference for the following discussion.

The intent for this figure is to differentiate between basic information, knowledge, and understanding. The fundamental input to the decision-making process is the information itself, that is, the fragments of measured phenomena or human inputs that represent some aspect of the real world and the desired end-state. Information alone, however, is not sufficient to support the decision. The information must first be assimilated and interpreted as knowledge, not simply as fragmented facts. Then the knowledge must be understood in terms of its implications on the ability to achieve desired objectives or the validity of current plans.

It is important to recognize that the decision-maker needs to process information inputs through several levels to filter, aggregate, correlate, and interpret it before it is useful. The model indicates that knowledge can be produced by acting upon the information inputs in a way that reflects the context of the situation. This is relatively straightforward (as discussed below) and lends itself to significant support from automated systems. The model also indicates that the conversion of knowledge to understanding and applying it in a decision involves further application of experienced-based judgment to the process. This is where the use of automation becomes more difficult and where the decision maker's personal qualities, expertise, and intuition become important. The providors of C4I information must understand and take advantage of information management and presentation technologies to help in both processes and to configure systems and their knowledge products to best support the decision makers in applying that experience, judgment, and intuition, which must always be part of the command process.

Further, information that is collected from the environment that feeds the process will always be fragmentary (never complete) and the perception of that environment will always be based on judgment rather than certainty. He or she will almost never know "ground truth" and only have a perception of the true situation based on the fragments and experienced-based judgment. Uncertainty will always be part of command; the role of C41 systems is to provide commanders with the best opportunity to recognize and deal with the situation, even though their knowledge will be incomplete.

4.2 Developing Knowledge: The First Step Toward Understanding

As previously discussed, the process for developing knowledge can be conceived in several distinct stages which are relatively straightforward and methodical, The first is the acquisition, filtering and conditioning of input information to assure that it provides needed knowledge, representing significant changes or verifications of the real world situation and not unnecessary or irrelevant information. The information must also present the added knowledge in a form that suitable for rapid recognition and application (without further study),

The next step is to combine information with the user's current knowledge of the situation in order to clarify, update, or confirm that knowledge. At this point, the inputs are subjected to further filtering and correlation with the user's "current knowledge" base. The information then must be projected into a "perception model" that the user employs to represent his or her environment or situation. The new inputs drive the model to update current knowledge and to verify whether the model outputs remain consistent with information from the known world. In many instances, the decision maker will have several alternative perception models, especially when uncertainty is large. As knowledge improves or as the patterns of knowledge are more clearly understood, the models may be necked down to one or a few, and the decision maker moves closer to being able to exercise judgment or intuition with confidence,

The first step to achieving an effective conversion of raw information to knowledge is to import the information in a form that is easy to use. In some cases, the form of the input can be used to directly convey significant amounts of knowledge, even before it is further processed. An example is the use of graphic or imagery transmission to convey a sense of the situation. Experience tells us that we prefer visual inputs to textual or numerical ones in most cases we can convey significant information by showing pictures, and we can do so without saturating the user. The saying "one picture is worth a thousand words" is actually quite close to the actual case. Consider, for example, the following comparison between the rates at which information can be "absorbed" by a user if we present it in visual versus textual forms.

Text can be read at speeds on the order of teletype transmission, 100 words per minute A fast reader can probably exceed this by a factor of three to five, but probably not ten, The standard operational vocabulary may consist of a few thousand words, so each word could be represented by a data block about 10 bits long. As a result, the 100 words per minute can be translated into roughly 1000 bits per minute of "word-oriented" information. English is a relatively redundant language, and narratives tend to mix real information content with words that satisfy grammatical and syntax requirements. Of the 1000 bits per minute, one-fourth or less are probably real information bits, even in a tightly constructed report. result is an effective information rate on the order of a few bits per second. This means that narrative input can overload the user with reading chores at an input information rate of only a few bits per second. Anyone who opens his or her e-mail in the morning to find fifteen or twenty new messages understands the burden of servicing textual inputs, The people who cope best are those who implement filters, such as profilers, to sort important messages from less important ones.

By comparison, imagery probably can be viewed and understood at several hundreds of bits per second of real information A reasonably good, full motion television picture can be processed, compressed, and transmitted at several hundred kilobits per second. Within that picture, only the changes are being sent. Human beings viewing this picture will normally focus on those parts that draw their attention, either because of prior understanding of where the significant information is in the picture or because of the rate of change in parts of the picture. Within the observer's field of interest, most of the information is background or context information, even if it is changing rapidly. For example, if an automobile is moving across the screen, the pixels that form the image of the automobile will be transmitted as updated information, even though most of those pixels define the shape, color, type of automobile, and so forth. They are transmitted because the image is moving and because they are part of that image, not because the details that they represent are changing in form or substance. Consequently, the actual information contained 'in the transmission will be a significant factor less than the data rate. The moving automobile is a good case in point. Here, we are probably viewing no more than a few tens of bits per second of real information input. If the automobile were not the only changing aspect in the picture, and if occupants were also being viewed, this could double or triple the information rate to the viewer. If much more is actually happening in the picture, the viewer will probably not be able to notice all the real information. Therefore, an overload limit for video presentation is probably on the order of hundreds of bits per second of real information.

Figure 3 illustrates a notional set of "information overload" or "user saturation" curves that might apply to the form in which the information is delivered. The shaded areas indicate where overload degrades effectiveness. This figure expresses the commonly accepted view (discussed above) that the form of presentation can improve both the effectiveness of the information and the degree to which large quantities of information can be accepted and used by decision-makers. Narrative information (generally teletype or electronic mail) tends to be relatively burdensome to read and to translate into data or other forms of presentation that support further processing into the knowledge base. Formatted information, either tabular or structured data, can be tailored more suitably for rapid insertion into briefings, summary displays, or files. Graphic displays are immediately understandable and can integrate significant amounts of information into concise pictures with overlay symbology or annotation.

Knowledge production involves other functions in addition to tailoring the product to the user's needs. Those functions consist of correlation (and fusion) of different pieces of information that are related and correlation of the information with the context and the prior knowledge. When these are completed, the decision maker is presented a "knowledge product" that usually will consist of some pattern of information, arranged to convey a clear message. Knowledge is closely coupled to the decision maker's perception of the environment or situation. Here we begin to build understanding for the appropriate decisions to be considered. Up to this point, the processing of information has been relatively straightforward. Inputs have been filtered for significance and quality. They have been aggregated and correlated according to types. They have been compared with preexisting knowledge, and they have been organized and displayed to match the desired view for decision making. Each of these processes is relatively mechanical or algorithmic in nature. Hence, automated systems that exist today are quite good at accomplishing them. The result is usually some form of presentation that appears on a desktop terminal or large screen display and that provides the human being with a piece of the puzzle,

The real challenge is to help the human being place that piece correctly and recognize the puzzle pattern when only a fraction of the total pieces are available. The human being may be an analyst, a staff officer preparing a briefing, or the commander who must make a decision. In each case, the human being needs to be able to take the knowledge products and to use them to enhance his or her perceptions, This is where technology is starting to have significant impacts and where C2 doctrine organizations, concepts, and procedures must be reengineered to exploit and leverage emerging technologies.

4.3 Building Understanding

We now consider the use of the knowledge products to build understanding. Knowledge is useful only if it can be understood in terms of the implications for the decision. If the situation is relatively simple or if it does not change rapidly, the interpretation of knowledge in terms of decision criteria can be relatively straightforward. As complexity, dynamics, or uncertainty increase, the use of the knowledge can become a burdensome and labor intensive process.

Reasoning support is an area that is growing in importance as commanders address increasingly complex and time-sensitive decisions and have large amounts of supporting information to assist them, and where significant uncertainty may also remain. The principal resource available to the commander for perceiving the situation and understanding it is his or her experience and judgment. If a certain pattern of information has been encountered previously and always represented a clearly defined situation, the commander will likely recognize that pattern and make the connection quickly. A less experienced person might need more information, or might reach a wrong conclusion. This is part of the process for training and practicing, to become an expert in the field of endeavor. A number of technology areas are also particularly relevant to building understanding and are important adjuncts to the formal training and practice necessary to develop expertise.

Expert systems and artificial intelligence are obvious enablers for reasoning support. Their designations are deliberate references to human qualities that signify expertise and good judgment. These types of tools attempt to provide a decision-maker with an ability to quickly appraise information and select among alternative courses of action to respond. Most of these automated aids attempt to mimic the heuristic, intuitive decision making that the experienced and proficient human decision maker would exhibit.

Expert systems provide an automated aid to make the selections based on a wide variety of experts who have been confronted with accumulated experience in similar situations. Artificial intelligence can assist the decision-maker by identifying features and factors that may not have been obvious and by suggesting linkages to other information that may be useful in making the decisions. Artificial intelligence may also be useful in providing an initial basis for understanding partial information, by filling in the blanks with likely estimates of missing information. If the system is truly trusted, it may even be used to provide the summary assessments and recommendations that the decision-maker will act upon directly.

Object-oriented technology is emerging as one of the very powerful factors in providing a means to integrate among the information, knowledge, and reasoning products presented to the decision-maker. Object-oriented information processing has a virtue, among others, of being able to present and interact with complex information and functionality in a relatively simple way, shielding the human being from the underlying complexity and heterogeneity of interconnected (feeder) systems. It also provides a means to establish automated functions that are matched to real world objects and activities, rather than application programs that are coded to represent predefined processes. This offers flexibility to structure the human computer interaction to match the decision maker's actual needs rather than the predefined concept of how he or she will operate.

Figure 4 illustrates a concept where the decision-maker can view a presentation and is able to call for additional information related to the presentation. The example illustrates a hypermedia type of process that allows the user to select specific "pointers" to types of information packaged within the delivered product. This is indicated by the dashed lines and shaded areas in the presentation and the linked information files. The user draws information from those linked sources with the aid of knowledge access tools or "robots" (called knobots) that are integrated into the knowledge product. The user may not be directly aware of where the information resides or how it was linked to the delivered knowledge product, Current applications available for use on the Internet can provide some basic types of access, retrieval, and in-depth search capabilities.

More advanced capabilities are emerging as part of object-oriented computing architectures and should become important features of information systems of the dawning 21^st Century. The example in Figure 4 can be thought of as representing a "knowledge object" that includes the summary presentation and the linkages and rules for accessing all information and applications related to the knowledge object, This will be discussed later as part of the reasoning-support construct.

Consider the following hypothetical example. A commander is presented with a new "knowledge object" representing an enemy mobile command center that had not been observed in the area prior to this report. What the commander sees is a visual representation of the object, maybe a cartoon or an icon with some annotated alphanumeric information, but the object also contains a number of linkages and functionalities that it carries with it,

The commander queries the object to determine why the system identified this as a command center. The query may be a simple "click of the mouse" and the stroke of one or two keys. Linkages to the source data show the commander that specific direct observations plus some unspecified intelligence led to that conclusion and advise the commander that confidence is greater than 95%. The commander then queries the knowledge object to obtain information regarding the implications on his perception of the enemy disposition and likely intent. The object contains additional linkages to order of battle, technical characteristics, and enemy doctrine and tactics files. These may be embedded directly in the object or they may be achieved by secondary linkages that the object has to other objects or applications, but that are invisible to the human being using the system. Additional linkages may be contained to expert system applications that can estimate likely size and composition of forces to which this object is normally attached and to systems that may provide some indication of enemy intent, given the known presence of the object. The main point is that a query may cause a very complex set of searches to be exercised, and a summary report is presented to the commander. The result may be a revised estimate of the tactical situation with confidence of 80%. The commander may choose to respond by adjusting his or her own plans. The example given above is a forecast of the type of automated support that may be realized in the coming decade. Reasoning support at a somewhat lower level of capability and complexity is already being used in some cases where relatively straightforward expert system applications are currently available. These cases include weaponeering, mission planning, deployment planning, force module selection, communications planning, and mobility planning.

As expert systems and other artificial intelligence and simulation tools (or knobots) become available, the use of reasoning support is likely to grow significantly and to become an important part of C4I system core capabilities.

Now we get to very bottom-line observations on the manner in which the evolution (or re-engineering) of command and control can produce a paradigm shift (or revolution) in our national security capabilities. Emerging technology must be exploited and leveraged while adapting it and applicable command and control processes to significantly improve the effectiveness of decisions at reduced costs. First, let us examine the drivers for decision making and distinguish real drivers from artificial ones that enables us to see more clearly where current practices can be improved in areas that are driven by legacy concepts and procedures that are sub-optimal in the context of their potential in a new paradigm; then we can examine some of the key technology thrusts that offer real increases in value in achieving enhancements that we seek.

5.1 Enabling Technology

As noted previously, the real impacts on command and control will probably come from changes in doctrine, organization, facilities, concepts, procedures and related training and education to take advantage of new technology. Revolutionary enhancements come only when we change the way we operate (technologies are only enablers). If we do not take advantage and leverage the enablers and make the changes, our effectiveness increases only incrementally (we merely pave the cowpaths), If we exploit and leverage the enablers, we can make revolutionary improvements. Command and control decision support is the one operational area where significant opportunities are available for revolutionary change. To capitalize on those opportunities we must first identify the real drivers for C2 decision making. There are mainly two of these that cause the commander to make a decision and act upon it.

The first is the situation itself. At certain points in time, the commander must make a decision because the operational situation demands it. This imposes a clear and hard limit upon the decision making cycle, regardless of the availability of sufficient information, knowledge, or understanding.

The second is the degree of knowledge and understanding of the commander himself. At a certain stage, when he or she has a sufficient set of knowledge and understanding, the commander will recognize the opportunity for a decision and that the supporting information is sufficient for the decision. This may occur even when the operational situation is not forcing a decision. It may be a self-generated understanding that a decision can improve some aspect of the operation. A highly skilled and practiced commander will reach this point earlier than a less skilled one, relying on past experience, judgment, and intuition to recognize an important decision opportunity. Note that we use the word "opportunity" as opposed to the previous notion of "demand."

We must apply technology to support the commander in reaching decisions as opportunities rather than demands. That is, the military commanders should be provided the capability to operate and make decisions as the knowledge and understanding facilitate and stimulate those decisions. We must decrease the degree to which he or she is in a reactive mode, making obligatory decisions before he or she would choose to voluntarily.

The third type of decision driver that we need to recognize is an artificial one, even though it is one of the principal drivers in today's command centers. This is the procedure-driven decision. If the headquarters procedure is to develop a briefing for the commander every twelve hours, the staff will develop their inputs according to that schedule. They will make numerous assessments and decisions as part of the staffing process and the commander will be presented with some decision-oriented parts to the overall briefing, This structured process is an established way of doing business in a headquarters, largely because of the manual nature of the work and the need to coordinate and integrate among numerous staff. Without well defined, consistent procedures, the process would become chaotic. Nonetheless, this decision driver is actually artificial as it relates to the existing environment or situation,

If technology is implemented with a view toward the real decision drivers, and if the doctrine, organization, facilities, concepts, procedures and training and education are adapted as well to address them, major improvements are possible. This can only be accomplished through close coordination between operators, developers, and educators in managing the necessary changes based upon an underlying re-engineering of basic processes. With this in mind, let us identify some of the major areas where technology holds promise for dramatic improvements command and control and ultimately in affordable national defense capabilities.

5.2 Information Exchange Technologies

A readily apparent way of improving the delivery of information and knowledge to the decision maker is to disseminate information in a form that is easily managed, presented and quickly understood. In many cases, the best dissemination medium is visual and the best presentation is an image or graphic. Traditional military communications are extensions from telephone and teletype transmission that have been mainstays for command and control for the better part of this century. Information conveyed by these means is primarily narrative and serial in nature. As noted previously, this limits the real information rate to several bits per second and limits the presentation of knowledge to relatively narrow domains of information that are received sequentially. They must be correlated with one another and integrated by the user, placing a burden on him or her to manage the inputs and determine whether and how they fit together. They do not provide easily recognized and understood patterns of information. Emerging technology in the electronic-mail category (more recently with multimedia attachments to datagrams) promises to greatly improve the effectiveness of information transport. For example, current procedures cause an analyst to look at a map or image and compose a textual message telling the recipient something about the situation. The recipient will read the message, interpret it, and then reconstruct the original picture that caused the message to be sent. Needless to say, this is laborious, time consuming, and error prone. New technical capabilities can provide attachment of the original information, the picture or map or table, so that the recipient can have the information in context in an understandable form in a timely fashion. This provides the capability to send data that can also be directly integrated into the picture; thereby, the commander has multiple sources of inputs which can be rapidly integrated into that picture in parallel and almost simultaneously. Not "magic" or "futuristic" thinking, it is currently available capability. Internet services and similar defense and national networks at each level of security provide this type of multimedia information transport, and common desktop computers can access and use the information in powerful ways.

For example, the evolving Defense Message System (DMS) will greatly improve previous types. of messaging and common e-mail types of services into a modern unified enterprise-wide messaging service. DMS will allow multimedia attachments that will be matched to the standard desktop environment as well as the automated C41 infrastructure. Operators will have an exceptional opportunity during the next few years to demonstrate and evaluate these new tools and to adapt their own procedures and operational concepts to use them effectively. Movement toward "smart push" as well as "smart user pull" of information is a significant aspect of the emerging way to significant improvements in information dissemination management. The ability of the user to select the information needed (and when) is closely tied to the movement away from traditional text messaging and toward multimedia messaging as discussed above. The DMS and related enhancements to the defense information infrastructure are enabling technologies which will allow the kind of flexibility and responsiveness needed to make this practical.

5.3 Building Knowledge Products

In the past, the decision maker and supporting staff had no choice but to take individual pieces of information and assemble them manually to form understandable pictures, such as a situation plot. Automated processing has made major advances in performing this function. A number of systems currently exist for handling information and constructing products that can present highly correlated and integrated knowledge. The Global Command and Control System (GCCS) is available and emerging unified capabilities that support greatly improved applications that can be tailored to users' needs.

The most basic capabilities are those that apply deterministic filters, such as geographic or type filters, and algorithmic correlation processes to consolidate relevant information from multiple sources. Most modern command and control systems employ these underlying support capabilities. More sophisticated correlators also exist for combining dissimilar data related to the same object or event and to create a picture of that object or event based on the full set of measured information. This is sometimes referred to as fusion rather than correlation, due to the combining of dissimilar types of information to form a new class of knowledge regarding the object or event, A number of intelligence systems and the GCCS apply this type of capability to some degree, especially to fuse data from sensors (such as radar) with data from intelligence sensors and other sources. This has led to the current concept of C4I systems linked to surveillance and reconnaisance systems collectively referred to as "C4ISR" systems. Tremendous opportunity exists to improve the decision maker's knowledge of the situation by building further upon existing and emerging technologies. But many current systems are adapted from paradigms of the past, in which specific types of information are isolated (stovepiped) into different parts of the staff. In such cases, information such as force status, intelligence, weather, and logistics are still handled as separate "knowledge domains" and the overall integration must be done either in the decision makers mind or with multiple displays or "windows." Emerging enabling technology allows the decision maker to bring information together from across multiple "knowledge domains" to clarify a situation or display a problem, For example, weather or terrain information can be integrated with situational information and operational planning information to give a more complete picture. Similarly, force deployment and movement plans can be integrated into the situation display to give a picture of how and where the arriving forces or supplies will integrate into the battle plan. The defense supporting establishment must do a better job of working with the military operational professionals to explore opportunities for adapting technology to their needs and also to allow them to explore changes to their concepts, doctrine, organization, facilities, procedures and related training and education to exploit the potential offered by technology. It is clear that the greatest enhancements will come from adjusting processes to take advantage of enabling technologies. Simply applying them to the current processes will not provide the leap ahead in achievable defense capabilities which are urgently needed in the uncertain international landscape in which we find ourselves, with fewer resources at our disposal.

5.4 Tailoring Knowledge-Based Information

The previously mentioned capabilities for improved information dissemination management and for knowledge integration are enablers for the tailoring of knowledge presentation to suit users' needs and preferences,

At present, C2 systems are designed by engineers, based on their discussions with program managers and their judgments as to what the operators want and need. Subsequently, program managers are presented with information displays and input-output capabilities that were specified and designed by the engineers. Modem technology offers an alternative: flexibility and adaptability in design, allowing the users to custom tailor the "form, feel, functionality, and content" of the delivered knowledge products

A simple example is the selection of the "up-down" aspect of a display. In some cases, we prefer "North up" In other cases we prefer to see the operational situation with the opposing force at the top, the "forward line of troops (FLOT)" horizontally at the middle of the display, and own forces at the lower half of the display, In cases where we are managing movements, we often prefer "heading up" in the display. Other examples are easy to find, including preferences for symbology and overlays, background map features, and the like The point here is that the knowledge should be presented to the user in the form that is most easily understood by that user in coping with his or her situation.

Recall that decision makers recognize patterns of information which they can interpret based on their training and experience. They will recognize certain patterns and react to them based on judgment and intuition if and only if those patterns are presented in an effective form. They also must be able to interact with the knowledge products in ways that will provide amplifying information that their judgment recognizes as relevant. A simple picture may sometimes suffice, but more often the decision maker will want to know more about a specific aspect of that picture. The knowledge product must allow for this type of unstructured query enhancement.

Systems are now emerging which provide effective knowledge presentations, with the needed interactive query capabilities. Once again, GCCS is an excellent example of such a system. Among other applications, GCCS enables a common integrated situational display and linked query capability to underlying data behind the symbols in the display. Additional "hypermedia" presentations, extending to complex "web objects" are part of the emerging automated processing environment. The challenge now is to form a closer relationship between the operators and the technologists so that these features can be inserted into C2 systems in an adaptable manner suitable for users' needs. We must also continue to help operators recognize how evolving capabilities can be exploited and leveraged to adapt concepts, doctrine, organizations, facilities, training and education and procedures Thereby, we can obtain much more than small "deltas" for marginal improvements.

5.5 Knowledge Robotics (KNOWBOTS)

Reasoning support, whether with expert systems, simulations, artificial intelligence, or some other new approach, is probably the most exciting new capability on the technological horizon.

While the machine cannot now (or in the foreseeable future) take the place of the human decision maker, it can provide an enormous enhancement to the human's ability to see the implications of current knowledge and potential evolutions. In the simplest case, the machine can perform a number of "look ahead" projections to help evaluate alternative courses of action. These can be straightforward, algorithmic calculations of resource expenditure or mobility factors. They can be more complex statistical analyses of attrition rates or other aspects of combat. They can even apply heuristic reasoning to project outcomes and compare alternatives, When the decision maker asks the "what if' questions, machines can provide some part of the answers.

Three types of reasoning support seem very attractive for near-term applications in improving decision making without certainly. One is the "expert system," the second is "probabilistic analysis," and the third is "simulation and wargaming." These are not the only automated aids that support reasoning and decision making, but they are important enough to warrant individual attention.

Expert systems attempt to share behavioral experiences of experts with decision makers, providing them the benefit of the experts' experience-based judgment. The trick here is to capture the "picture" that the expert actually saw, identify the pattern the expert reacted to, and define the heuristic relationship between the "cause" (i.e. the recognized knowledge and level of uncertainty) and the "effect" (i,e. the expert's decision). Since the expert's decision was probably intuitive and judgmental, the automated system will exhibit some degree of apparent "learning" even though it is simply capturing lessons from human experts and correlating them. In some cases, this can provide a powerful adjunct to the human decision maker, since it can quickly cut through confusion or irrelevancies and focus on specific important issues. It can also provide good advice on reasonable courses of action to consider and direct the focus on these as opposed to other (less promising) courses. A major challenge here is to inspire the military professionals' confidence in such systems and to have them contribute their expertise to the data bases.

Though intuitive decision making is recognized as most prevalent by experts or masters of a craft, analysis tools still have value. Probabilistic analysis is recognized as a traditional analyst's tool, however, the use of probabilistic analysis to guide complex, heuristic decision making is still in its early stages for command and control processes. Several examples exist where C2 systems make use of probabilistic approaches to guide decisions, particularly through the use of maximum likelihood estimates to support surveillance and search operations. Other applications assist with transportation and mobility and to project likely actions of an adversary. The use of probabilistic analysis with operational planning applications offers a potentially significant advantage to operators in evaluating plans and intended actions. Again, the challenge is to bring these new technologies out of the laboratory and into a representative operational environment so that they can be adapted to the operators' needs in a manner (and at a rate) consistent with building confidence and blending them into the human decision making processes,

Simulations are probably the easiest automated aids for military operators to accept as tools of their trade. This is because the simulations are easily recognized in terms of the specific questions that are being addressed. Simulations represent objects and events that the decision makers understand, and the outcomes are usually easily understood as well. Since there is less mystery surrounding a simulation, at least in terms of the visibility that the user has, there is greater acceptance of simulation as a tool for making projections. A danger here is the users' incomplete understanding of assumptions on which the simulation model is based and its limitations. Simulations are becoming more important as embedded tools within the systems, due primarily to emphasis by the R&D community and enthusiasm by operators to obtain these capabilities. The challenge is to improve the speed with which "what if' analyses can be accomplished and to integrate the simulations more thoroughly into operational systems and the overall C2 decision making context, including operators' procedures. That leads to the next offering of enhancement opportunities, the use of technology to improve the proficiency of the decision makers and their staff.

5.6 Technology-Assisted Experiential Learning

The final area of opportunity is the application of automated aids for training and practice in C2 decision making. LTGEN Van Riper, USMC (Ret) has emphasized that commanders and staffs require extensive practice to develop a level of expertise, judgment and intuition to make decisions quickly and precisely, even in the face of uncertainty. He points out that the major applicable human cognitive ability is the recognition of patterns of knowledge and the correlation of those patterns with judgmental or intuitive understanding of the situation and alternatives courses of action. Human beings, unlike machines, do not simply process all the input information according to some algorithmic process and derive the "answer" based on deterministic analysis. As stated previously, human beings recognize patterns and determine their responses based on meaning recognized from those patterns, based on past experience or judgment that is derived from a broad variety of experience and insight, This is one reason the military commander cannot easily be replaced by a machine. The machine has its place as an aid in decision making but lacks the judgment and intuition to take actions based on heuristic appraisals,

The challenge to the defense supporting establishment is to provide current and future commanders and staffs with realistic mechanisms to develop and hone their judgment and intuition. Clearly, real world problem solving is the best way to do this, but real-world military actions are not everyday occurrences. We must provide some way for command and control professionals to develop the required skills and to practice them sufficiently to develop the judgment and decision making style suited to command in modem warfare,

Current technology offers a variety of embedded tools to enable this, and the advent of interactive gaming and simulation will expand the offered capabilities There are basically three levels from which to address the problem of training and practice.

The first is the basic level at which tutorial drills are embedded in an automated system to teach users about its individual features and functions. This is a very rudimentary (but important) part of most automated systems.

The second level is the embedded training software that provides representative work for the user to accomplish, using the automated tool as part of his or her work process. This emphasizes using the tool as part of the operators standard work processes and procedures This is important as a way to cause the operator to incorporate the tool as part of the process in a natural way,

At the third level we reach the point where the tool becomes an important part of the decision makers cognitive processes, not just a mechanical adjunct to a predetermined process. This requires that the user begin to adapt both the tool and his or her own thought processes to a common, integrated approach for developing and recognizing patterns of information. The tool can become an important part of the judgmental decision making process only if the user understands its full capabilities and is comfortable with the degree to which he or she can use the tool to reach decisions. In order to achieve this level of integration of the automated systems and the intuitive decision making process, the users must have an opportunity to practice extensively with their tools in very realistic situations, including all the uncertainty and stress that would be reflected in real world situations. This means that the C2 systems would have embedded within them very realistic operational problems with very realistic interactive responses. The analogy can be made with a realistic video game played on the C2 terminals, similar to the simulators used in training air crews. Such embedded training is an essential Ingredient of future C2 systems. At this point the "system" links the operator with the equipment and software/

The challenge to the C41 community is to evolve improved capabilities for military commanders to cope with uncertainty and make decisions based on sound experience-based judgment and intuition, We must provide support to the decision makers from a wide variety of information resources (organic, regional and world-wide). We must also provide improved capabilities for information dissemination management to get at information that is crucial to the decision at hand. Commanders also need systems that are "smart" enough to recognize when to 'push" information to a decision maker even if he or she has not asked for it.

Equally important is increased emphasis on C41 automation as a means for supporting intuitive decision making processes through the effective presentation of knowledge and understanding. This is in contrast to generally applying automation to analytical constructs. This means better dialog between operators and developers to adjust and-adapt their efforts to improve command and control processes and capabilities. Generally, we must recognize that automation is becoming increasingly part of the decision making process itself, not just a set of tools that used within traditional processes of antiquated paradigms. Most important, however, is the need to maintain an understanding of command and control decision making as a function that deals with uncertainty. It is not a straightforward, algorithmic process whereby the decision is "calculated" as a numerically precise result. It involves human decision makers who will confront problems under considerable stress and uncertainty and who will have to make decisions from experience-based judgment. They must be provided the means to extract the best understanding of the situation based on available information and to extend this understanding based on intuition and perception. Possibly, in the distant future, we will have developed machines that can match the human being's capability for pattern recognition and intuitive decision making; but until then, the human decision maker is still the key element 'in the "loop" and information technology must be configured in such a way as to optimize his or her effectiveness within that loop,

Athens, Major Arthur J, USMC - "Unraveling the Mystery of Battlefield Coup d'oeil," a monograph at the School of Advanced Military Studies, U.S. Army Command and General Staff College, Ft. Leavenworth, KS, February 1993

Klein, Gary A. - Naturalistic Decision Making: Implications for Design. Wright-Patterson AFB, Ohio: Crew System Ergonomics Information Analysis Center, April 1993.

Marsh, Howard S. - "From the Fog of War to Information Overload; A New Challenge for Command and Control," the MITRE Corporation, 1994

Schmidt, John F. - "A Concept of Command and Control," Fleet Marine Force

Reference Publication (FMFRP) 15-3, Headquarters, U.S. Marine Corps, 3 August 1994

Van Creveld, Martin *- Command in War. Cambridge, MA: Harvard University Press, 985.

Van Riper, Paul K. "Preparing the Marine Corps for Combat Operations in the 21stCentury: The Command and Control Challenge," prepared for the National Security Studies Program, (NSST-578-03) Security Decision Making, University of Georgetown, Washington, DC, May 2, 1994