3
Committee Member Observations on Adapting Additional Simulation Techniques for the Air Force
AREAS WHERE THE AIR FORCE COULD BENEFIT FROM ALTERNATIVE USES AND TECHNOLOGIES
A plethora of observations resulted from the presentations in Chapter 2. This section begins with extensive observations regarding future simulation architectures (Box 3-1, Box 3-2, and Figure 3-1, with associated explanation).
Committee member observations touched on the broad concepts above. First, Don Fraser, committee co-chair, and several other committee members were optimistic that, based on the earlier presentations, a significant part of this architecture concept is already in place (e.g., in the distributed mission operations network known as DMON). These committee members noted that movement forward can thus evolve in stepwise fashion with advances sized to meet specific training needs. (Note: Col Nathan Hill, Chief of ACC Flight Operations, mentioned issues in this area: “How many networks are too many? What type of networks do we need? What are the second and third order effects of shutting down and consolidating networks?”) John-Paul Clarke opined that the Air Force needs a modular-flexible framework as a strategy on which to hang tactics and mechanisms to promote convergence versus a large program of record. He went on to say that it is necessary to know what standards to use. John Corley offered that the development of an intellectual architecture for live, virtual, constructive (LVC) simulation must occur prior to contracting for the physical architecture. Mr. Corley supported the use of the Drew-Robinson architecture concept. Mr. Corley believes the intellectual construct should not demand investment but provide a framework for decision makers to “opt in” where LVC supports learning opportunities not available through other methods, or where value is enhanced. He stated that the system design must be sufficiently adaptive to delivery of knowledge that, on the whole, delivers learning that is more rapidly assimilated and retained for longer periods.1
Committee co-chair Ray Johns offered that having established standards will allow the Air Force to have lower life-cycle costs. In a related topic, Michael Zyda noted that the U.S. government has failed miserably in simulator standards. He said, “Why not use open source procedures and processes?” Ex-post facto standards are hard to do, and very expensive. Dr. Zyda stated that the National Research Council’s 1997 report Modeling and Simulation: Linking Entertainment and Defense, which he chaired, raised almost all the same issues with respect to the internetworking of defense simulations.2 The lengthy architecture dialog led to the following additional key themes.
_________________
1 Bimal Aponso, Chief, Aerospace Simulation Research and Development Branch, a suggested using a phased approach to developing the common architecture using limited operational scenarios. The stated aim of this approach is to reduce the risk of integrating LVC components. Large-scale demonstrations and tests are inherently difficult to assess in terms of effectiveness due to the sheer scale of the variables involved. A phased build up to a large scale test using smaller, easier to measure, operationally relevant scenarios may be a better approach.
2 For additional information, see National Research Council. Modeling and Simulation: Linking Entertainment and Defense, Washington, D.C.: The National Academies Press, 1997.
BOX 3-1
Observations on Path Forward for Integrating Air Force LVC Efforts
Pamela Drew, Committee Member
1. The implementations of live, virtual, constructive (LVC) simulation for training currently underway in the Air Force, Navy, and elsewhere are being developed in independent, stovepipe, and ad hoc fashion, which results in a platform-centric capability with simulator-simulator (hardwired) interfaces, disintegrated networks, and duplicative and similar, but unstandardized and unshared, data and mission sets. An alternative, and what is needed, is an approach that creates a common architectural approach in which LVC simulations can be “plugged” into an integration LVC backbone or integration architecture—hereafter referred to as ILVC-IA. Figure 3-1, from co-member Harry Robinson, illustrates this type of architecture.
a. In this architecture, there would reside reusable data for terrain, weather, threat information, blue tasking, etc. It would also contain reusable mission models, mission logic and rules, and simulators that could be re-purposed and used in various applications or instances of ILVC training sessions. The live or VC simulations would be integrated into this environment via standardized interfaces for communications and data links, for SIM via DIS and HLA, and the data passed would have to conform to standardized access interface protocols. Using this common integration architecture and enforcement of standards, a proprietary solution can still be integrated as long as it conforms to the interface and data access requirements.
b. This architecture can be put into use to support the entire range of desired combinations of LVC to support all missions from the “high end” Combat Air Force (CAF) requirements to more routine VC training scenarios. These mission scenarios create use cases of the architecture and results in specific applications or instances (e.g., an F-35 Live pulling VC world view of KC-46, AWACS, weather, etc.).
c. Of note, real-world sensors can also be integrated as feeds into the system, thereby bringing “reality” to the simulation. Obvious examples are for terrain and weather as part of the “live” feed, as well as other live assets.
d. Finally, security was referenced in multiple ways as a gap or obstacle by various presenters. In the ILVC-IA, security would have to be addressed. A few different elements would include encryption for the transport layer; multilevel security for crossing classification levels; role-based, access-control-type capability for authentication and authorization; and physical security for facilities.
2. By creating this new architecture, it would be possible to transform from a platform-centric view to a reality-centric view, enable more rapid integration of simulated and live assets, and enable far more efficient development of training capabilities.
Theme 3. Currently, live, virtual, and constructive training efforts are evolving in a largely ad hoc, stovepiped, and somewhat inefficient fashion. This situation suggests Air Force consideration of a different architectural approach that would be world-centric—open, pluggable and playable—rather than platform and contractor proprietary centric. This world-centric construct would contain common elements and live data, such as weather, terrain, threats, with an array of specific simulation platforms around the periphery drawing information from the common databases as opposed to utilizing their own proprietary database (Pamela Drew, Harry Robinson).
Theme 4. There are indications that some elements of the Air Force simulation architecture currently have these world-centric enterprise characteristics, so continued pursuit of an enterprise-level solution to live, virtual, constructive training could be very beneficial (Pamela Drew, Harry Robinson).
3. While this can be viewed as a technical architecture, the Air Force sponsors see it as providing a framework to articulate potential investment needs and to prioritize “where the next dollar should be spent.” CAF and Mobility Air Forces (MAF) representatives both commented that, of the data sets presented during our general discussion, geographic, terrain, and threat sets were the priority.
4. CAF has an emergent and urgent need to bring VC simulation to augment F-35 Live to enable training due to constraints stated in the workshop. These are a combination of the decision not to allow full capability in live training, amongst others.
5. There is a need to organize the development of such an architecture through a clear authority structure, which would lead the architecture, standards, interface, and reusable asset-data-capability development. Note the goal should be to leverage all that can be reused or adapted to that which already exists.
6. The advanced technology demonstration (ATD) presented by Wink Bennett (Air Force Research Laboratory, AFRL) is an example of one “bottoms-up” instance of LVC underway. This could be harnessed and adapted as needed to become a first instance to begin implementation of the ILVC-IA architecture.
7. Just as such an architecture would benefit the Air Force, there is an analogous gap and application across the services—Department of Defense (DoD) wide. The Navy is also just beginning the LVC journey, developing yet another (mostly separate) capability operating on the JBUS, which appears to be the counterpart to the Air Force distributed mission operations network (DMON). Getting the services to use the Defense Information Systems Agency Global Information Grid (DISA GIG) via the Joint Information Environment (JIE) will facilitate the transport/network layer of integration.
8. There are various efforts underway that address some or perhaps all of the proposed ILVC-IA. These include the J7’s JLVC Vision 2020, the AFRL ATD, Air Force Special Operations Command Ops training, and industry efforts (e.g., Boeing, Lockheed, Northrop capability). These efforts should be assessed and leveraged into this unified ILVC-IA capability as possible and appropriate.
9. There is a need for a single authority within the Air Force to define architecture, enforce standards, to select and maintain reusable content of the ILVC-IA, including, but not limited to, reusable data and mission sets. The authority should also create and drive execution against a near-, mid-, and long-term roadmap and associated plan that demonstrates capacity to integrate legacy capabilities (both government and industry) with new capabilities. In addition, and as important, are a new governance model, communication model, and stakeholder engagement.
10. There were a variety of technology developments and improvements for human-in- the-loop interfaces (e.g., Google Glass) and techniques (e.g., motion) that can be included in a continuous technology refresh sub-task in the oversight and development of the ILVC-IA. These assessments must also be specific to training objectives.
Committee members had additional observations in other areas. First, Robert Allardice noted that, in connection with the Boeing presentation, a benefit is that current and emerging technology for assessment and gaming technology may provide significant growth in our understanding of learning. He also noted that mobile technology has changed how people make decisions; we ought to heavily leverage mobile technology for enhancing learning and substituting training. Mr. Allardice went on to say that more discussion should take place on what the Air Force understands about “how” humans can best learn “today” based on significant discoveries and advances within the past decade. He said it is important to tailor the right learning tool for the right learning objective and place competency in the right platform. Finally, Mr. Allardice shared that, regarding the medical presentations, there are tremendous lessons to learn from collaborating with the medical community. That community, he noted, is advancing understanding of learning, education, and training, making significant advances in several technologies that could help the Air Force prioritize and match content to learning platforms. Leveraging technology to deliver an experiential learning environment similar to medical simulation is important.
FIGURE 3-1 Notional architecture for U.S. Air Force live, virtual, constructive training.
BOX 3-2
Explanation of Notional Architecture for U.S. Air Force LVC Training
Harry Robinson, National Program Manager, Veterans Health Administration (VHA) Simulation Learning Education and Research Network (SimLEARN) (Committee Member)
Simulation-based training environments for the Air Force would benefit from an architecture using Common Reference Access “Bus” that would serve as a shared information provider simultaneously supporting generation of mission characteristics and events necessary to provide realistic training. Components would contain grouped databases that would drive LVC simulations accessed from training platforms unique to specific aircraft types, models, and series. Each component database would be established and subsequently maintained to achieve necessary level of currency. The respective modules could be characterized as Unclassified (including terrain, weather, navigational aids, air traffic control, and white force generation) or Classified (including threat, enemy orders of battle, common sensors [similar across multiple aircraft], common weapons [air-to-air, air-to-ground, similar across multiple aircraft], command and control, communications, data links, blue force generator, and intelligence-surveillance-reconnaissance scenario injects. Simulation execution of unique platform models for aerodynamic performance, aircrew interface (e.g., controls and displays), weapons, and sensors would integrate with the components’ data accessible on the common reference access “bus” as controlled or limited by a multilevel security filter. Specific aircraft simulators would plug in to the common bus. Advantages of this construct include (1) reduction in need for stove-piped and proprietary solutions for each type aircraft simulator, (2) standardized component databases that can be independently established, (3) ease for maintaining database currency, (4) networked simulations executed in a shared environment, and (5) simulation-based exercises that support specific platform security program requirements.
John Corley offered that the ability to deliver learning for training is important and that consideration should be given to the changes in how airmen will “learn” and the related future demographics. (Note: Col Hill also had a comment in this area: “The new generation learns differently than most of us. We need to figure out the best way to teach them.”) Steven Detro observed that, as a substitute for some training, higher-fidelity simulation technologies are now enabling more training to be accomplished in virtual reality. He went on to note that continued analysis of the potential benefits that virtual reality simulation could offer to each area of training should be considered. With a blend of different training media and training devices, Mr. Detro offered that a greater percentage of training sorties or training events could move into simulators; these hours should complement current live fly hours. (Note: Current acquisition policies have forced the Air Force to find lower-cost technology.) Mr. Detro believes the Air Force could use (1) performance measurement technologies already developed by the Air Force Research Laboratory (AFRL) to increase the ability to objectively measure the effectiveness of training;3 (2) the “science of learning” cognitive modeling products of the AFRL to assist in the development of more efficient learning delivery methods for training; and (3) immersive technology advancements to deliver training information to match the learning preferences of students.
Several committee members provided final thoughts in this area. First, Ray Johns offered that independent research and development by industry has advanced knowledge of simulation applications and technologies. Second, Harry Robinson noted that, to meet Air Force needs, the following outside capabilities and technologies are most useful: adaptive learning and intelligent tutoring, cloud computing, common accessed resources for data, real-time representations and feedback loops that avoid latency issues, and multiplayer interactive gaming that builds teamwork and communication skills. Lastly, Michael Zyda noted that, with respect to the CAE presentation, open standards for all parts of the simulation enterprise will decrease costs and make better systems. He said that, regarding the Lockheed Martin presentation, alternate simulation systems become possible with head-mounted displays, and perhaps the Air Force should look at head-mounted displays and augmented reality technology for some of what it is doing. He also thinks that networked simulators have latency problems; perhaps look at what the game industry does for this.4
HOW LIVE, VIRTUAL, CONSTRUCTIVE TECHNIQUES COULD IMPROVE AIRCREW TRAINING
Several committee member comments applied to how techniques for LVC simulation for training could improve aircrew training, although there are links back to other messages in this report. For example, Robert Allardice noted that at one point simulator training was secondary; however, advances in technology have led to LVC as the primary way to train for the mission. Mr. Allardice went on to say that he thinks the best way to frame LVC is not that it will improve training. It is that advances in technology and modern applications drive an LVC “imperative.” Mr. Allardice noted that all training can benefit to some extent; the key seems to be to develop an architecture from which specific applications can draw, based on the risk profile the Air Force chooses based on a particular mission set. John Corley was of the opinion that the Air Force needs to make prudent investments that enable needed enhancements to or development of the enterprise intended to yield a “realistic” training environment. Mr. Corley noted that
_________________
3 There was a comment from an operational pilot in the audience on the need to measure training effectiveness. The pilot said that when developing a training capability, particularly of threat environments, it was important to ensure that the probability of success in the simulator be equivalent to that in an actual situation. This highlights the overall issue of ensuring a viable training effectiveness validation method is developed in tandem with LVC simulation capability.
4 For best practices, consider the following publication: Johns Hopkins University Applied Physics Laboratory, Best Practices for the Development of Models and Simulations: Final Report, NSAD-R-2010-037, Laurel, Md., June 2010, available at http://www.msco.mil/MSBPD.html.
investments must consider the temporal dimension, bit-sized approach toward the delivery of the LVC capability. The following themes arose during the discussions:
Theme 5. Advances in technology and increasingly complex user needs have led to live, virtual, constructive training as the primary way to train for some missions (Robert Allardice).
Theme 6. Substantial benefits could accrue to the Air Force if it relied on open systems and acquired data rights as the model when procuring new systems. Enforcing compliance to more interoperable, related standards could lead to a “plug and play” environment (Pamela Drew, Michael Zyda).
Theme 7. Research into the “science of learning” is indicating that young people, who have considerable computer skills compared to previous generations, learn in very different ways compared to older generations. Future architectures and systems would benefit by taking this knowledge into account (adaptive learning) (Donald Fraser, Steve Detro).
SUGGESTED AREAS FOR POSSIBLE FOLLOW-ON STUDY
Ray Johns indicated that the Air Force sponsors of this workshop requested that there be no follow-on study. Nevertheless, some committee members suggested a few areas that the Air Force may wish to delve into more deeply; these areas are listed below.
1. What is the full set of requirements for Air Force LVC simulation for training?
2. What is the optimal standard and architecture that the Air Force should strive for? What is the roadmap for the architecture?
3. How can multilevel security be dealt with—through a study in its own right? and Should such a study be classified?
4. What can be done about adaptive learning?
5. What is the need, if any, for a change in Air Force governance with respect to LVC simulation for training? What organizational and budget changes need to be made for an effective LVC simulation for training capability across all missions (with the F-35 as the first system priority)?
This page intentionally left blank.
