The Tactical Simulation System (TACSIM) provides interactive computer-based simulation to support intelligence training from MI Battalion through Echelons Above Corps in exercises such as REFORGER, Central Fortress, Ulchi Focus Lens, Team Spirit, Warfighter, and others across Germany, Korea, and the United States. In near real time mode, TACSIM aids in the training of intelligence staff skills from the design of collection requirements to the analysis of raw intelligence. TACSIM models the tasking, collecting, and reporting functions of specific U.S. reconnaissance assets. The system architecture and interconnections to other simulations are shown in figure 1.
The model was originally developed as the Post Oak Simulator System (POSSIM) in 1979 under the Army Training and Doctrine Command (TRADOC). In 1980 it was renamed TACSIM and one of its first missions was to stimulate the All Source Analysis System (ASAS) with intelligence reports created in realistic format and volume. From a scripted scenario database, TACSIM would operate intelligence missions over enemy forces and generate reports in United States Message Text Format (USMTF). These would then be provided to ASAS at multiple classification levels forming the perceived picture of a conflict in progress.
Given the standard format and realistic volume of TACSIM output, it was a natural candidate for stressing intelligence analysts involved in training exercises. Prior to its introduction into the training arena, intelligence analysts were often trained by human scripters who were responsible for replicating the performance of collection assets and writing the appropriate reports. Other intelligence simulations were/are available, but none contain validated sensor and platform models of the fidelity found in TACSIM. Another significant characteristic is the fact that TACSIM produces standardized USMTF reports, enabling the model to transmit simulated intelligence directly into the same assets that process real-world intelligence. The message parsers of systems such as Warrior, Hawkeye, the Electronic Processing and Dissemination System (EPDS), Enhanced Tactical Users terminal (ETUT), and Constant Source can process TACSIM reports without having to make modifications and work-arounds. This translates into more realistic training by using the same systems the way they will be used in the actual combat - supporting the philosophy "We Train Like We Fight".
The TACSIM suite of equipment includes a standard Communication Support Processor (CSP) for transmitting data to exercise participants at multiple security levels. This real-world communication connection enables TACSIM reports to be transmitted across AUTODIN and local transmission lines. Therefore, since most intelligence units have access to AUTODIN they can participate in exercises held around the world, further supporting the "We Train Like We Fight" paradigm.
The TACSIM system contains Army and Air Force validated models of selected intelligence collection assets. The distributed wargaming architectures emerging under the Distributed Interactive Simulation (DIS) and Aggregate Level Simulation Protocol (ALSP) architectures will provide a foundation for TACSIM to support any exercise in the world from any TACSIM equipped simulation center. This means that less equipment and fewer personnel need to be deployed to support international training objectives, greatly reducing operational costs and risks.
Communications Intelligence (COMINT) has traditionally been very difficult to simulate without the intervention of human scripters to provide realism. TACSIM has developed a method of providing realistic COMINT that corresponds to the dynamic characteristics of the exercise through time. This model provides a much more realistic intelligence product than has been available in the past, with the goal being to eliminate the COMINT burden from human scripters for most common combat situations.
Each COMINT report is capable of describing communications between commanders and subordinates, subordinates and commanders, lateral echelons, and echelon skipping. The information in a report can lead to knowledge about the location, activity, and identification of both the transmitting and receiving units.
Within the report is a section for recording conversational narrative. In the past this narrative has been selected from a large library of scripted sentences and phrases. These were grouped according to the type and activity of the unit whose communications were intercepted, providing a minimum level of correlation between the scripted conversation and events occurring in the exercise. This concept has been greatly improved and expanded to provide COMINT that is better correlated to exercise events. Since the software has instant access to a huge amount of simulation data, it is better informed than any human scripter could ever be.
Scripted conversations are now categorized according to many different unit characteristics, making it possible to generate very detailed narratives that will be used only under specific conditions. These characteristics include the type, echelon, and activity of the detected unit. In addition, keying on the amount of time the unit has been engaged in its current activity results in narrative reports that correspond to the doctrinal operations that a unit performs as it establishes defensive positions, goes on the offensive, or begins to convoy. Since the narrative library has to be created prior to the exercise and may be used for multiple exercises, it is not possible to include details about exercise events and unit identifiers at that time. To enhance the realism of these narratives, TPO has developed a method by which the model can dynamically include the unit name, type, echelon, activity, location, and call sign in the narratives as they are being generated during the exercise.
The level of detail provided by the TACSIM COMINT model has freed human scripters from many laborious and mundane tasks and allowed them to focus on special cases and high priority events. As a result exercise participants are now receiving computer generated intelligence which far surpasses any that has been available during the past decade.
The sensor models in TACSIM have been tightly controlled to assure accurate portrayal of the actual sensor platform and system reporting capabili- ties. Recent requirements have dictated that, in addition to traditional sensor suites, TACSIM must accommodate conceptual and newly developed assets. Generic sensors and platforms have been added that can be customized to replicate any definable system that provides imagery or electronic intelligence. This is especially useful for assets such as unmanned aerial vehicles that can be configured in many different ways for an exercise. Platforms can be designed with a wide variety of operating characteristics and limits. Characteristics such as range, altitude, velocity, and banking angle are used to describe the platform. Each of these platforms can carry multiple sensors packages, each with its own physical, electronic, and imaging operating characteristics. Units will now be able to test the contributions of new and proposed assets in exercises as soon as this type of performance data is available.
Recent requirements of training exercises have resulted in the need to allow TACSIM product to be disseminated to Allied forces such as Korea and NATO countries. Since the model operates at multiple security levels, some of the raw product is not releasable to these forces. It is also common for these countries to send command staffs to exercises without the aid of their intelligence analysts. The volume of data provided by TACSIM would overwhelm a small contingent and require them to perform analysis which is not their function. To meet this new requirement, the TACSIM Project Office (TPO) developed the TACSIM Analysis Operations Node (TALON).
TALON (Figure 3) is designed to perform intelligence analysis as it would be done by trained human analysts. TALON is inserted between the player and his TACSIM data connection and can handle any message product that the exercise participants cannot receive directly. All intelligence reports are then processed and a NATO/ROK releasable summary produced. The reports are in the form of Situation Reports (SITREP's) and Size, Activity, Location, Unit, Time, and Equipment reports (SALUTE's). These USMTF messages are supplemented by textual statistics messages describing the type and volume of TACSIM reports that went into producing the summaries. TALON reports are disseminated to exercise participants through the same communications system used by TACSIM. The system has also proven useful at exercises in which U.S. forces do not have access to secured facilities or do not have the support of their intelligence analysis elements. These players can receive summarized collateral product even though they are tasking a mixture of COMINT, ELINT, and IMINT sensors.
The heart of TALON is the tailoring and analysis module, composed of algorithms broken into four primary groups. The first extracts and scores data from newly received reports. The second correlates the new information with existing intelligence on the detected unit to form a single perception of the enemy situation. The third translates numerical decision coefficients into english text. Finally, the information is prepared for inclusion into a USMTF message for transmission to exercise participants.
The data produced by TACSIM and TALON becomes voluminous during an exercise, in some cases exceeding 100,000 intelligence reports. The ability to review this information in an analytical manner to find successes achieved and mistakes made is essential to fully benefit from the training scenario. In order to support this need, TPO provides the TACSIM After Action Review User System (TAARUS). This is a graphical query system with access to all intelligence data that has been stored in a relational database.
TAARUS receives a copy, in real-time, of all reports transmitted to every exercise participant by TACSIM and a copy of the exercise ground truth data. Using a window-based interface, an analyst can access any piece of information moments after it has been distributed to the training audience. Data can be grouped and filtered by characteristics such as time, location, intelligence discipline, exercise participant, sensor mission, and a host of others. The output can be formatted as maps using Defense Mapping Agency terrain and standard military symbols, multi-dimensional graphs, predefined priority information windows, and tabular data generated by user defined queries. Samples of this output are shown in figure 4.
Using this system, exercise controllers can be kept abreast of the latest intelligence events without having to task an entire staff of people to gather and collate the data. Map and graphical formats enable the presentation of large amounts of data in a very usable form that can be further supported by the extended detail found in the tabular reports and summary windows.
The mission of the National Wargaming System (NWARS) is the simulation of tasking, reporting, and the dissemination of information from national intelligence collection assets for the purpose of training and exercise support. Developed by the Defense Support Projects Office (DSPO), these models include electronic, electro-optical, synthetic aperture radar, and infra-red sensor collection. NWARS reports are standard USMTF and can be released through the CSP just as those from TACSIM are. NWARS is integrated with TACSIM such that both share the same input and output connections to the exercise combat driver, but it can also operate in a stand-alone mode with direct interface to exercise participants. In this case mission tasking can be entered directly or remotely through collection management devices such as SWIFTHAWK.
The scenario databases that drive CBS, TACSIM, NWARS, AWSIM, etc. are extremely large, detailed, and complex. Preparing and formatting this information prior to an exercise has typically been a very man-power intensive task, requiring many man-months of effort. In order to minimize this cost, DSPO and TPO have developed the Rapid Scenario Preparation Unit for Intelligence (RASPUTIN). RASPUTIN is capable of constructing an accurate database of friendly and enemy forces using detailed knowledge bases of doctrinal procedures, force structure, fixed target characteristics, equipment characteristics, terrain limitations, and weather data.
With this automated support RASPUTIN operators can construct an entire exercise scenario in a matter of weeks rather than months. Once constructed the scenario can be formatted for insertion into many different simulations (e.g. CBS and TACSIM), guaranteeing that all systems will begin the exercise with completely compatible representations of the battlefield. The graphical interface and multi-simulation compatibility have been estimated to cut an 18 man-month preparation effort to 2 man-weeks.
Once a scenario has been detailed according to RASPUTIN's knowledge bases, it is possible for operators to customize the data according to their specific situations and objectives. This provides the advantage of both doctrinal exactitude and human experience in determining the final exercise database.
[The simulations described in this article are directed by the U.S. Army Simulation, Training, and Instrumentation Command (STRICOM) TACSIM Project Office, 7900 Sudley Road, Suite 500, Manassas, Virginia 22110, 703-830-7607, and by the Defense Support Projects Office, 703-222-7634.]