ADDING AN INTELLIGENT TUTORING SYSTEM TO AN EXISTING TRAINING SIMULATION Richard H. Stottler and Randy Jensen Stottler Henke Associates, Inc. San Mateo, California Bill Pike STRICOM Orlando, Florida Rick Bingham MÄK Technologies Cambridge, MA 02138 Battle Command 2010 (BC2010) is a tactical decision game used by Command Prep Course students at the Command General Staff College at Fort Leavenworth to play battalion level tactical scenarios in a dynamic, 3-D environment. The use of this simulation, however, still required the effort of an instructor to observe the student's actions and provide an after action review (AAR). It was determined that the addition of an Intelligent Tutoring System (ITS) to BC2010 would off-load the instructor from these duties and allow the students to execute scenarios without requiring an instructor for the AAR. This paper presents the lessons learned from that experience. In BC2010, students playing a scenario must first read the mission background which includes the mission objectives and five paragraph order. They then develop a plan and input that plan into each unit under their control. They monitor the execution of their plan and the tactical situation in 2-D and 3-D views. Enemy units are only shown when they are sighted by friendly units. During the simulation the students can issue real-time commands. The ITS is interfaced to BC2010 via the High Level Architecture (HLA.). Initially the student's plans are transmitted from BC2010 through HLA to the ITS, before simulation execution begins. These plans are critiqued by the ITS by comparing them to good and common bad plans for the scenario, as determined by a subject matter expert. The student receives this feedback and corrects the plan. Execution then begins. BC2010, through HLA, sends to the ITS both the locations and actions of vehicles and the commands sent by the student. The ITS evaluates the correctness of these actions, given the current circumstances, determines which tactical principles the student has correctly applied and which have been missed, and automatically assembles a debriefing. It can then recommend further study and additional scenarios to improve the student's weakest areas. Biographic Sketches: Dick Stottler co-founded Stottler Henke Associates, Inc. (SHAI), an artificial intelligence consulting firm in San Mateo, California in 1988. He has been principal investigator on a number of tactical decision-making intelligent tutoring system projects conducted by SHAI and is working on an ITS for battalion commanders at the Command General Staff College and a prototype for the future combat system. He has a master’s degree in computer science from Stanford University. Bill Pike is the Lead Principal Investigator for Advanced Distributed Learning (ADL) with the US Army STRICOM. At STRICOM, he has been the principal investigator on several ADL topics on integrating courseware with advanced topics, including the use of game engines as assessment tools for tactical principles. He is an officer in the Naval Reserves and has a Master’s degree from the UCF in Computer Engineering. Mr. Richard Bingham is the Director of Programs at MÄK Technologies where he is responsible for the daily operations of all government contracts relating to MÄK's SIMinterNET, PC-based training applications business. Before joining MÄK, Mr. Bingham worked for Sikorsky Aircraft. As a simulation engineer he was responsibly for the physics based modeling of helicopter dynamic components and Mission Equipment Package components. Randy Jensen is a Project Manager for Stottler Henke Associates, Inc. (SHAI) and has developed intelligent tutoring systems and prototypes for a variety of domains, including current projects for battalion commanders and personnel at the Air Force's Aerospace Operations Center. He has a BS in Symbolic Systems from Stanford University. Report Documentation Page Form Approved OMB No. 0704-0188 Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. 1. REPORT DATE 3. DATES COVERED 2006 2. REPORT TYPE 00-00-2006 to 00-00-2006 4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER Adding an Intelligent Tutoring System to an Existing Training 5b. GRANT NUMBER Simulation 5c. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION Stottler Henke Associates Inc,951 Mariner’s Island Blvd Suite 300,San REPORT NUMBER Mateo,CA,94404 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S) 11. SPONSOR/MONITOR’S REPORT NUMBER(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for public release; distribution unlimited 13. SUPPLEMENTARY NOTES The original document contains color images. 14. ABSTRACT 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF ABSTRACT OF PAGES RESPONSIBLE PERSON a. REPORT b. ABSTRACT c. THIS PAGE 10 unclassified unclassified unclassified Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std Z39-18 ADDING AN INTELLIGENT TUTORING SYSTEM TO AN EXISTING TRAINING SIMULATION Dick Stottler Stottler Henke Associates, Inc. San Mateo, California Bill Pike STRICOM Orlando, Florida Rick Bingham MÄK Technologies Cambridge, MA 02138 Randy Jensen Stottler Henke Associates, Inc. San Mateo, California PROBLEM DESCRIPTION needed to monitor the soldier’s actions and decisions, recognize the critical teaching points, coach higher levels of performance, and provide remedial instruction. Instructors and experts agree that the most important factor in developing skilled tactical decision-making is BC2010 DESCRIPTION practice making tactical decisions in tactical scenarios. The Department of Defense (DoD) has long recognized this as well. They first addressed it primarily through Battle Command 2010 (BC2010) was developed for the field exercises. These were expensive and available Battle Command Battle Lab at Fort Leavenworth, only a small fraction of the time. This problem was Kansas. Part of the Intermediate Desktop Trainer (IDT) addressed through the development of simulations. But genre of simulations, BC2010 provides an environment these were originally large, expensive, and costly to for Army Commanders and their staffs to evaluate skills operate. in planning and executing tactical operations from a brigade level and below. The system has been designed In the last five years, much progress has been made in to allow both single-player and multi-player operation developing low-cost simulations to support military for collaborative execution and can be used for head-to- training and education. Likewise, the military head game play. Using standard desktop personal recognizes the potential of advanced distance learning computers, the trainer has an embedded simulation and is striving to maximize its use for education and engine that is capable of modeling opposing red forces training. These efforts have, in many cases, reduced the and adjacent blue forces. Additionally, a built-in after- overhead in computers and manpower needed to train action-review (AAR) system is available to record the our soldiers. Today, soldiers, using a single personal complete networked mission, playback the mission on a computer, can practice their tactical skills in a 2-D map or 3-D environment, and create statistical simulated battlefield environment. Only a few years charts for analysis. ago, this same training would have required multiple computers, several computer operators, observer- BC2010, in its simplest form, is played in a single controllers, and an exercise coordinator. Battle player mode. In this mode, the commander is Command 2010 (BC2010) is a tactical decision making responsible for planning all battlefield functional areas game. It is an example of how a low-cost simulation (BFAs) and during execution, commanding all units. can be used to enhance advanced distance learning. The trainer, however, is more commonly played with several users over a network. In this mode, one player BC2010 represents the next generation of tactics traditionally assumes the role of the brigade trainers. However, one hurdle remains before the commander. The other players, during planning, power of the PC-based tactical simulation can be assume the roles of the staff officers (e.g., Maneuver, harnessed: the incorporation of an intelligent tactical Fire Support, Engineering, Intelligence) and during tutoring system. Until this is done, an instructor is execution, the unit commanders. Once a mission is selected, both planning and game positions and support by fire positions, all graphically play are performed through an intuitive game-like user represented with associated locations. The player can interface, shown in Figure 1. BC2010 has two primary then use the graphics on the overlays (e.g., a planned displays, the 3-D “pop the hatch” stealth view and the minefield) as objects in the simulation on which 2-D map view. The 3-D view is used so the player can subordinate units can be assigned to perform tasks. view the outside world and obtain first hand knowledge Thus, all simulated units can be assigned plans to be of the outlying terrain. This view provides a realistic performed during execution. representation of the world. The 2-D map view is used for all planning and execution-based activities. During execution, the user shifts gears and is quickly engaged in an ever-changing battle. The user utilizes During planning, the map view permits the user to the map view to monitor progress of the troops and graphically plan a BFA and view the plans from other make modifications to units’ plans as required. BFA’s through the use of multiple tactical overlays. Complex, multi-phased plans can be modified as new These overlays, shared over the network in a multi- enemy intelligence is gathered, or instant action player game, represent the acetates that would be used commands can be issued to units as time becomes more by the staff officers in the tactical operations center critical. In multi-player missions, communications (TOC). Objects placed on the overlay are used to during execution are performed through the use of a portray the intentions of how the player will perform text-based chat or a voice-over-IP system, permitting that portion of the operation. For instance, a maneuver collaboration in the tactical decisions to be made. staff officer might place several routes leading to battle Figure 1. Battle Command 2010 Game Interface ITS CONCEPT DESCRIPTION taught, and to decide on the next appropriate simulated scenario for the student to run. Intelligent tutoring systems (ITS) can best be defined as Intelligent tutoring systems are ideal for incorporating advanced training software that mimics a human tutor desktop free-play simulators into computer-based by adapting its instructional approach to each individual training since the software can stand in for a human student. They are particularly valuable for teaching tutor in all the roles. Existing IMI course material can complex cognitive tasks such as trouble shooting, often be integrated with ITS-enabled simulator and problem solving, and resolving critical situations. As a other active training. In this way, ITS technology can human tutor does, an ITS continually monitors and greatly leverage the training value of existing IMI and assesses each student's actions, infers the student's state desktop simulators. As shown in Figure 2, the ITS can of knowledge, and decides on the next instructional monitor a student's interaction with both simulation and event to maximize the student's learning. To do this in a other training content in the IMI, create and update the significant and cost-effective way, intelligent tutoring student model and decide on the next instructional systems use artificial intelligence. event (e.g., provide a hint, ask a question, run a new scenario, display multimedia to explain a concept, alert One-on-one tutoring by skilled human tutors is widely a human instructor that the student needs special help, regarded as the single best mode of instruction. A study etc.). by Benjamin S. Bloom of the University of Chicago To keep track of each student, the intelligent tutoring and Northwestern University concluded that, under the system creates and maintains a "student model" for best learning conditions they could devise (tutoring each individual from the first time he or she logs onto one-on-one), the average student was 2 Sigma above the software. Depending on the sophistication of a the average control student taught under conventional particular intelligent tutoring system, the student model group methods of instruction. That is, the average keeps different amounts of information on the student. tutored student was above 98% of the students in the The most basic information includes the tasks the control class. student has performed as well as performance information on those tasks. From this information, the Conventional Interactive Multimedia Instruction (IMI) software estimates the student's mastery of relevant software is not designed to provide such a high level of skills and knowledge, and the student's ability to apply adaptive response to each individual student as an them when appropriate. For example, a student may be individual human tutor or ITS can. In fact, most IMI able to apply a concept in one set of circumstances, but software more closely resembles an "electronic not under other circumstances, so it is important that the textbook" rather than an "electronic teacher." Just as a software tests the student's knowledge of each concept book implicitly encourages a student to start at the front under different circumstances. and move to the back, such IMI software usually encourages a student to move linearly through a set of multimedia material, with the occasional multichoice ITS questions to test the student's retention of the information. Such tests do not assess the student's ability to apply the information. The ability to apply information in a job should be the goal of training. Scenarios Also, conventional IMI is not able to meaningfully Monitor incorporate use of free-play simulators into their curriculum. This is a major shortcoming of Simulation IMI conventional IMI as student manipulation of sophisticated simulators that realistically replicate issues that they will encounter on the job is widely recognized to be a highly effective training technique. The catch has been that simulators without instructors are virtually useless for training, and their unsupervised use can even result in negative training. Students working on simulators need instructors to point out Student their correct and incorrect actions, to brief them in Figure 2. ITS can integrate free-play simulators and context on the underlying concepts that are being IMI BC2010 ITS DESCRIPTION thoroughly into the BC2010 application, as shown in Figure 3. Overview The integration of an existing simulation with an Figure 3 illustrates the interaction between BC2010, ITS, existing Intelligent Tutoring System has several and the student in the ultimate configuration. During the advantages. planning phase, the student obtains preliminary plan information from the courseware and develops a detailed • Leverage of familiar tactical training tool: Both plan in BC2010. Once the plan is complete, the ITS students and instructors at CGSC are familiar with evaluates the plan and provides feedback through the BC2010. In addition, CGSC plans on using BC2010 BC2010 interface to the student. This is provided to the as their simulation tool during a pilot program in student through annotated tactical overlays and animations 2002. that illustrate the likely outcomes of a bad plan. The student is able (and encouraged) to modify the plan, and then • Leverage of existing intelligent tutoring system: moves to the execution phase. During execution, the STRICOM leveraged their investment in the existing student receives both 2-D and 3-D data from the BC2010 ITS technology as opposed to developing an interface and provides command and control information to intelligent tutoring system from scratch. The two the BC2010 simulation engine. In the background, contractors (the ITS developer and the developer of BC2010 sends the ITS the internal data necessary for BC2010) did need to work closely to ensure analysis. By monitoring the student actions in the simulated successful integration of the ITS with BC2010. scenario, the ITS assesses their correctness in the current situation. The results are used to debrief the student by automatically assembling an ITS-based After Action Tactical Plan Evaluation Review (AAR). The AAR will ultimately be provided to Tactical plan evaluation requires the capability to the student through the BC2010 interface, and will use the evaluate two factors: (1) placement of the correct kinds current BC2010 AAR capability as the baseline. Currently, of graphical overlay elements in the correct locations the ITS uses its own interface for all debriefing. After the within a predefined tolerance, and (2) suitability of the student finishes the scenario, the ITS will infer the roles assigned for each unit in coordination with knowledge deficiencies of the student and formulate a overlay elements. Suppose for a simplified example, remedial instruction plan, which normally includes further that Figure 4 is a correct plan supplied by an expert. course material, examples, and further BC2010 exercises, The instructor also annotates the correct plan, in a based on Command Prep courseware, to practice and test separate file, with an overall description of the concept the student's weaknesses. of operations, the rationale for why that concept is a good solution, and the principles that the student must understand to have arrived at this plan. The annotation files also allow the instructor to create similar annotations for each individual symbol. These are descriptions, rationale, and principles for the five elements of the symbol - why the tactic represented by the symbol is needed, why the type of unit or maneuver was chosen, why the size of unit was chosen, why the general location was chosen (which usually relates to tactical considerations of the overall plan) and why the specific location was chosen (which usually relates to terrain features). All of these descriptions and rationale take the form of referenced multimedia files so that animations, graphics, and other multimedia can be used in the explanations, instead of simple text. Figure 3. Interaction Between BC2010, Intelligent Tutoring System, and Student Currently, the ITS acts as a separate application and provides feedback through its own user interface. The second integration effort will integrate the ITS more Figure 6 shows an incorrect plan that fails the second evaluation factor, the assignment of roles. Figure 4. Sample Correct Plan Supplied by Expert Ech1 and Ech2 represent two companies treated as echelons 1 and 2 for this example. Obj1 and Obj2 are Figure 6. Second Incorrect Plan as Entered by Student two objective regions and Rte1 and Rte2 are the appropriate routes for echelons 1 and 2 respectively. In this case, the student understands the correct routes The student is presented with the same scenario and any and objectives, but issues commands incorrectly, in the background information or intelligence, but without the sense that the wrong units are sent to the wrong route arrows. The objective areas may or may not be objectives, possibly presenting time-space-distance provided, depending on the nature of the scenario. For problems and also potential coordination problems as example, a trainee may be expected to determine on his units move across each other. own what the effective boundaries should be for the objective areas, given terrain features or other factors. Tactical Execution Evaluation All elements of the correct plan are represented in a The BC2010 environment presents free play simulation hidden layer of the overlay which is only viewable by of battlefield conditions, so the ITS evaluation of instructors or course developers. Figure 5 shows an student performance often depends on the observable incorrect plan that fails the first evaluation factor in two accomplishment of certain simulation states that ways. involve the relative positions, orientations, and activities of more than one coexisting simulation element. Finite State Machine (FSM) evaluations provide an effective means for monitoring simulation states and triggering analytical conclusions when certain conjunctive conditions are met. Figure 7shows an example of a simple scenario in which student performance is evaluated with respect to a conjoined set of simulation elements. Figure 5. First Incorrect Plan as Entered by Student Rte1 is incorrect because the end point or destination is outside of the effective area of Obj1. This kind of example may happen in cases where the student is required to determine where the objective area should effectively be, so in this case the student may have misread a terrain feature. Rte2 is incorrect even though it has the correct end point at Obj2, because the route Figure 7. Scenario with Conjoined Simulation Elements clearly does not match the correct route in the correct plan within any reasonable tolerance. This student In this scenario, we can imagine that the student is would receive in his planning debrief the instructor- tasked with blocking an approaching enemy unit, given entered multimedia rationale for the exact location for the intelligence report that the enemy battalion is Rte1's end point (perhaps that location represents a approaching on the road shown. In this case, Obj1 piece of key terrain) and the rationale for the general represents an area visible only to instructors, not to the and exact location for the correct Rte2. student. The definition of the boundaries for Obj1 becomes necessary for the evaluation of the student’s performance in terms of reaching an effective defensive position. So in this example, the evaluation engine compare the student's plans to all of the plans created checks for a simple set of conditions, using two by the instructor for the scenario and pick the closest functions – GetCurrentPosition, which returns the given matching one. It then assembles a debriefing based on entity's position, and PositionMatchesElement, which that one. For common student mistakes, instead of the checks if a given position matches with a given element rationale explaining why the plan's overall concept was from the overlay. This is illustrated graphically in chosen, it explains why the overall concept is bad. If Figure 8. the student matches a bad plan, in addition to the explanations as to why it is bad, the student will also get a description of a good one and why it is considered good. The plans and plan symbols also have principles to be passed or failed depending on whether or not the student's symbols match them. In this way, the process that assembles the debriefing (picking the most closely matching plan and comparing its symbols to those of the student's plan) is also used to assemble lists of which principles the student successfully applied in a mission planning context and which ones he could not. This is used in further remediation as described further below. Since BC2010 sends the ITS plan elements as they are created by the student, it is possible for the ITS to provide instruction during the planning process instead of waiting until the plan is complete. We've developed Figure 8. Finite State Machine Example ITSs that present this instruction in two ways. One is in the form of a Socratic dialog. That is, the ITS asks the This simple FSM reaches the “Pass” state if the student student general tactical principle questions particular to has successfully brought the unit designated as echelon the specific scenario and the partially completed plan. 1 to the objective 1 area before Enemy reaches Bridge. These prompt the student to think about tactical This transition would also send a message to the ITS. principles that appear to be lacking based on the plan so This message would include a debriefing message that far. The other type of instruction is generally termed the action was correct and why. This would have been coaching. Coaching provides hints to the student while previously attached to the transition by the instructor. he is developing his plan. The best hints are the ones These messages become the basis for briefing and real- that provide a minimum of information with little time coaching as described in the next section. specificity yet get the student to apply the appropriate tactical principles correctly in the planning decisions. Remediation This is generally accomplished by providing hints that are very general at first and then increasing their The ITS remediates the deficiencies it finds in several specificity as required to elicit a correct decision. Of ways. One of the most important is the debrief (also course the evaluation system has to be kept informed of called the after action review, or AAR) which the ITS the degree of hinting required for a student to make assembles automatically. There are two types, since the each decision. Ultimately, hinting must be withdrawn student interacts with the combined system in two as the student's mastery increases so that he does not phases - pre-mission planning and real-time mission become dependent on it. execution. Close integration of an ITS with a tactical simulation The pre-mission planning debrief, as described above, provides an especially valuable form of remediation is generated by assembling the proper multimedia during the planning debriefing. When the student has rationale explanations for the parts of the student's plan created a bad plan, that plan can be simulated in faster that did not match the correct plan. However, this than real-time so that the student can see the method only is applicable if the student's plan is unfortunate results of that plan without having to spend reasonably close to the instructor's. This usually means the time to execute it. that they are in at least rough agreement about the concept of operations. To alleviate this constraint, the This also illustrates the importance of debriefing the plan instructor can store several correct plans as well as several common incorrect ones. The ITS will first development before moving on to execution for both instructional and automatic execution evaluation reasons. Without a dedicated plan debriefing, the student who has a coaching is deemed appropriate, the best hint is the poor plan will merely go on to execute it, spending least specific one that allows the student to make the considerable time running the simulated scenario before correct decision and is only presented to a student who finally getting the AAR at the end of the simulation. Only would make the wrong decision without it. The latter is then will the student be informed of the problems with the handled well by the student model. If the student has a poor history with a principle and application of that plan, too long after he had completed it, in opposition to the principle is necessary to make the current correct instructional principle of immediate feedback. decision, it is likely that the student will make a poor Furthermore, the student will have spent a large amount of decision without a hint. Furthermore, a general hint of time with this poor plan, reinforcing his memory of the the form "Consider " along with name of the principle poor plan. If the scenario happened to go well in spite of (such as "The Importance of Key Terrain") can be the poor plan, which often can happen, the student will easily constructed without giving much away. In the have favorable memories of the planning mistakes. This is event the student still takes an inappropriate action, the especially true when compared to the relatively small very specific hint of "do the correct action because ..." amount of time the student will spend in the debriefing of may still be better instructionally than a wrong decision his poor plan. By debriefing the poor plan immediately and and the delayed feedback of the AAR. directing the student toward the development (and then execution) of a good plan, only the positive plan will be As described above, the process of assembling both reinforced. Finally, it is much easier for the ITS to types of debriefs also generates, for each scenario, lists accurately evaluate the student's performance when the of passed and failed principles. This allows the ITS to student is executing one of a few known good plans. look at a student's entire history with a principle and decide what level of mastery the student possesses and The real-time mission execution debriefing messages whether the student needs remediation with reference to are assembled as described above by the transitions in this principle. This is generally indicated by poor the Evaluation FSMs. The transitions also generate performance with respect to this principle in multiple lists of passed and failed principles. To create the scenarios so that this type of remediation occurs outside automatic after action review, the ITS gathers the of the specific scenarios in which the mistakes were debriefing messages, organizes them and writes a made. (Scenario specific remediation was already multimedia AAR file organizing the actions, generally given in the automatic AAR.) Depending on the type in chronological order. The correct actions are of student and the severity of the problems, the student generally indicated in green, and they are accompanied may be given a description of the principle, a detailed by the explanation as to why they were correct along description of the principle, examples of the application with the principles that the student must have been able of the principle in other scenarios, and hints when faced to apply in order to have performed this correct action. with this principle in future scenarios. All students For actions deemed incorrect, the action is generally having problems with a particular scenario, after colored red and is accompanied by an explanation as to remediation, would receive additional exercises that why the action was incorrect along with the principles require application of the principle both to prove that that the student was not able to successfully apply. they can now apply it in an operational scenario and to (Failure to take a correct action is a common type of force them to practice the areas in which they are the incorrect action.) The ITS also writes out important weakest. events that don't necessarily correspond to correct or incorrect action of the students but provide important BC2010/ITS HLA Interface Description information as to what the tactical situation was at that BC2010 was already HLA compliant before this effort time so that the AAR file is easier to follow. The ITS began; likewise, the ITS already had the ability to also assembles lists of passed and failed principles. generate an HLA log file from an HLA-compliant simulation run and analyze it. However, there was a The same information used to compile the AAR file can desire to make the interface real-time so that real-time also be used to provide a real-time coaching component instruction (e.g. coaching) could be performed. for the student's real-time mission execution decisions. Consequently the ITS's HLA logger was converted into Coaching during a simulated mission is a matter of an HLA listener. Through the standard HLA Real- instructional philosophy. Some would argue that a Time Platform Reference (RPR) Federation Object coaching component is both unrealistic and disruptive. Mode (FOM), the ITS immediately had access to However the alternative is to both allow the student to information adequate for real-time mission execution make a bad decision (or fail to make a good one) and to evaluation. Most importantly this included vehicle delay the feedback until the AAR when the student will positions, velocities, fire events, hit events, and indirect be informed of the poor decision. In the case where fire events including their type. However, the ITS also was tasked with evaluating a student's plan, which is While the ITS was being interfaced to BC2010 it was not normally transmitted as part of a standard HLA also undergoing development unrelated to the ITS. compliant tactical simulation. This additional This was both positive and negative. On the one hand, information was also transferred to the ITS from developers were already working on BC2010 and were BC2010 as described below. therefore also available to make changes required for the ITS interface and to answer questions from the ITS The BC2010 simulation environment provides a set of team and in general, coordinate the development of the controls for assembling plan information in a combined system. However it meant that the ITS team distributed setting. With potentially several users was forced to interface to and work with a simulation viewing the same scenario, a plan can be that was undergoing active development, a moving collaboratively defined and seen at each user’s station. target so to speak. A plan typically consists of graphical elements defined in an overlay for a given map, coupled with specific Results commands for specific units, which may either refer to graphical elements from the overlay or function as The result of the first stage of the integration effort is independent commands. An example of a referential that BC2010 and the ITS are interfaced through HLA. command would be MoveAlong(route), where route is B2010 and the ITS has a coordinated set of predefined the ID for a route graphically defined in the overlay. scenarios, so that any scenario that the student is using An independent command would be MoveTo(x,y,z). in BC2010 as part of the combined system is known to Each unit or echelon may have a separate plan the ITS, in the sense of having predefined good and bad consisting of several commands, either referentially plans and predefined evaluation FSMs for it. The ITS related to the overlay or independent, and possibly successfully receives the plan information from including triggers based on test conditions. BC2010 and debriefs the student on the plan in its own interface. During mission execution the ITS receives In BC2010, the same mechanism that is used to create the state of the simulated world and the student's pre-mission plans is also used to issue orders during actions and successfully evaluates those using its FSMs. real-time mission execution. Thus, once the ITS was Again this debriefing is given to the student through the adapted to read the BC2010 plans through HLA (as ITS's user interface. described below), it was also immediately able to see the orders that a student was issuing to the units that he Lessons Learned commanded. The real-time mission execution evaluation could also consider a student's orders directly, instead of only being able to examine their HLA can be used effectively to interface an ITS and effect in the movements and actions of the vehicles. tactical simulation. Furthermore, the RPR FOM can be extended to transmit additional, nonstandard Since both the graphical overlay elements and the information, such as plan overlays. echelon commands are issued in real-time, the BC2010 application has a standard procedure for distributing The ITS that was interfaced to BC2010 already existed this information via HLA to all user stations engaged in and was interfaced to a variety of products to make a the scenario. BC2010 uses the RPR-FOM to transmit logically complete system. This made the overall data via HLA, but for this planning application some system very unwieldy and impractical for training. extensions to the existing FOM were necessary in order There were significant ease of use, development, and to correctly provide plan information. fielding advantages to interfacing the ITS to a simulation product which represents one self-contained The DtDataInteraction class is a general class of the solution with all the needed capabilities in one software RPR FOM for transmitting data, so in the case of plan package. However, interfacing to a simulation under information, a DtDataInteraction object is published by development was difficult. It would have been optimal the BC2010 application via HLA. The ITS, acting as a to interface the ITS to the simulation after the federate, includes a listener that parses the enhancements were complete, if time allowed. It is DtDataInteraction object to determine if it contains plan important to have the ITS team involved earlier so they information; e.g., a new command for a given echelon can influence the design of the simulation and what or a new graphical element in the overlay. If so, it information will be available to the ITS from the extracts the appropriate information for plan evaluation simulation. purposes. Tactical instructors are comfortable generating scenarios with a limited number of likely good and bad