SITIS Topic Details |
||||||
| Proposals Accepted: | |
| Program: | SBIR |
| Topic Number: | A10-167 (Army) |
| Title: | Algorithms for Ground Soldier Based Simulations and Decision Support Applications | Research & Technical Areas: | Information Systems |
| Acquisition Program: | Objective: | To develop, and demonstrate methodologies and algorithms that enhance analysis driven, ground Soldier centric constructive simulations and provide proof of concept for small combat unit decision support applications.
| Description: | Modeling and simulation (M&S) provides an analytic environment to assess potential materiel solutions in an operational context. However, many current constructive simulations require analysts to script missions in extensive detail and may also require analysts to detail response behaviors to multiple conditions. The end result is that current simulations often require significant set-up to realistically represent the spectrum of actions that may occur in a tactical operating environment. This may limit their realism, analyst set-up time, the analyst’s capability to fully assess materiel solutions or the ability to reuse scenario elements. One reason for this extensive scripting is that not many algorithms exist that would allow constructive simulation agents to be automated enough to respond appropriately to changes in their situation while at the same time executing tasks that have only been defined with mission goals, initial/boundary conditions, and constraints. Also, our small combat units are lacking decision support applications that can support execution of their operational tasks, even while advertisements claim that there are now over 100,000 iPhone applications. There is potentially significant overlap between many algorithms needed for a constructive simulation focused on small unit operations and those algorithms that would be useful in operational decision support applications that could aid our ground Soldiers. In both cases the need to understand and represent important elements of the real world is a key element in the development of algorithms and is a critical component in validation. Route selection and emplacement of forces or sensors are examples of tasks that need to be executed in both the simulation and in real operations where algorithms can be developed to support execution of the task. For example, many of the factors that are important to the real world unit leader are also ones that, if utilized, would improve the path selection and movement algorithms for constructive simulations. There are common elements in the development of many simulation algorithms and those needed in decision support applications. In each case we would need to: identify the decision that we are trying to represent or support; identify the factors that are important; identify and develop methods of obtaining the needed data; develop user interfaces; develop methodologies and computer algorithms; and address platform integration issues. In both cases, highly relevant factors will include METT-TC elements (mission, enemy, terrain and weather, troops and support available, time available and civil considerations).
This effort would research, develop, and demonstrate methodologies and algorithms that enhance automation and execution of a variety of ground Soldier and small combat unit (platoon and below) tasks within constructive simulations and provide proof of concept for small combat unit decision support applications. There are opportunities in numerous areas, to include; locating forces to minimize exposure while maximizing fields of view and fire, assessment of threats and vulnerabilities, servicing of targets or locating an enemy shooter. Potential approaches should address the data, methodologies, algorithms, and validation audit trail. Proposals should identify how the proposed research will advance the current state of the art. The products should result in more intelligent and realistic unit behaviors and improved execution of tasks within the simulation, an improved analyst capability to assess materiel solutions, and support development of battlefield decision support applications.
| PHASE I: Phase I will provide a proposed concept for the generation of algorithms and methodologies that support automation of tasks and missions within the simulation with less involvement by the military analysts and provide a proof of concept to potentially transition to ground Soldier battlefield decision support applications. The focus will be on urban operations. This will include identifying a number of specific operationally relevant decisions and actions that could be at least partially automated within the simulation and/or supported by a decision support application. The sponsor will assist with this process. The sponsor will assist with identifying a set of small combat unit tasks, initial/boundary conditions, links to higher level mission goals, military constraints, operational considerations, and other supporting data elements that will support execution of this effort. Examples of composite tasks that require the execution of more elemental tasks include cordon and search, building clearing, and enemy reconnaissance.
The proposed concept should support an extensible and generalized solution that is computationally efficient within the constructive ground Soldier simulation or battlefield decision support aid. Any data needs and assumptions required by the concept to be compatible with a constructive ground Soldier simulation should be clearly outlined and explained.
Phase I will perform a proof of concept that describes how one proposed concept may be utilized within a ground Soldier battlefield decision support aid and implementation in a specific constructive ground Soldier simulation, such as the Infantry Warrior Simulation (IWARS). Metrics in phase I will include:
- the usefulness of the algorithm to supporting quality analysis across a range of situations,
- the applicability and utility of an initial methodology and algorithm to be implemented within a simulation and/or decision application,
- the degree that it represents the important elements of the real world (valid),
- the applicability of the selected approach to the development of other algorithms,
- documentation and its ability to be demonstrated.
| PHASE II: Phase II will include identification and prioritization of additional operationally relevant tasks to include specific decisions and actions that lend themselves to automation through algorithm development. Phase II will also include design and implementation of multiple algorithms and methodology necessary in accordance with the Phase I concept. Sufficient knowledge elicitation will be conducted with small combat unit SMEs to ensure that critical real world factors (i.e. METT-TC) are identified and included in such a way as to support the development of each methodology, to include the necessary data elements and data structures. In Phase II, a plan will be developed to validate and test the algorithms and methodologies. The plan may also include how specific applications could be developed. A set of use cases that describe relevant military operations or missions could be utilized or developed to guide research and the methodology development. The use cases would also support testing and exercising the methodology and techniques. For example, one use case may be a scenario where the goal is to perform a cordon and search and the mission is complicated by the presence of enemy snipers that cause our forces to shift their locations while they respond. Development will lead to demonstration of algorithms developed in phase II within a simulation tool, such as the IWARS.
Other tasks include documenting and delivering a report including all algorithms, methodologies, and any data structures or software products necessary to support transition of the work to DoD simulation developers. The phase II report should also demonstrate and document how algorithms may be adapted or transitioned to support implementation into battlefield decision support applications. Metrics for this effort will include the number of methodologies and algorithms developed, the degree to which they represent the important elements of the real world, their potential utility within the simulation and/or decision application, documentation and their ability to be demonstrated. Other considerations include the degree to which the algorithms are computationally efficient, can be modified if additional elements need to be included, and can be implemented within a simulation or decision application.
| PHASE III: The developed methodologies and associated implementation have commercial applications in simulation products that incorporate decision-making and Human Behavior Representation. There may also be follow on work related to the continued development and transition of algorithms and methodologies for constructive simulations, e.g. IWARS or OneSAF. In addition, there is potential application of these simulation products to proposed DoD materiel solutions whose goal is to provide Soldiers decision support applications that will support enhanced Soldier situational awareness and improved decision-making. Also, there is potential application to future automated or semi-automated ground Soldier battlefield systems; such as a system that automatically distributes ammunition from a supply point when troops are in contact with the enemy. Non military applications could relate to security details, crowd control, or other operations where there are common elements with small combat unit operations.
| References: | 1. IWARS fact sheet, http://nsrdec.natick.army.mil/media/fact/techprog/IWARS.pdf 2. FM 7-8, Infantry Rifle Platoon and Squad, http://www.globalsecurity.org/military/library/policy/army/fm/7-8/ 3. Briefing provided to the Behavior Representation in Models and Simulation (BRIMS) Conference on 15 April 2008, entitled Link Between Human Behavior Representation for Models and Military Systems 4. Killzone AI: Dynamic Procedural Combat Tactics, http://www.cgf-ai.com/docs/straatman_remco_killzone_ai.pdf 5. Department of the Army Pamphlet 5-11, Verification, Validation, and Accreditation of Army Models and Simulations, 30 September 1999 |
| Keywords: | Decision Support Application, Soldier Analysis, Ground Soldier, Modeling and Simulation, Situational Awareness, Simulation Algorithm, Infantry Warrior Simulation. |
Additional Information, Corrections, References, etc.. |
Ref #3: Ref. 3: IWARS Briefing, BRIMS Conference, 15 April 2008, 12 slides. A10_167 Ref NSRDEC Auer BRIMS 2008 Final.ppt |
Questions and Answers: |
Q: Where can other resources, such as budget templates, formatting guidelines, etc. be found? |
A: Please refer to the SBIR 10.3 Solicitation which includes detailed instructions and information about preparing and submitting your proposal. |
Q: 1. Will you be providing a copy of the IWARS User's Manual to interested parties as part of this proposal process? |
A: A1. We are working to make the IWARS users manual available via a download from the SITIS website. |
Q: First person shooter simulations have been given a lot of development attention in recent years, primarily in the commercial world, and control on individual agents is already fairly sophisticated. Is the focus of this SBIR, then, intended to be more in the development of an overall, high level strategic command agent – something that can direct groups of agents to accomplish complex goals or change mission objectives in reaction to “the fog of war” – in other words, a General; or is the primary focus of this SBIR meant to create alternate means to control individual agents? |
A: The focus of this effort is not to develop an overall, high level strategic command agent or to control individual agents. |
Q: Is the focus of this SBIR limited strictly to controlling virtual war fighters in tactical simulations, or should the algorithms developed also be able to control civilian and/or hostage elements within a simulation? |
A: The focus of this effort is not to control the virtual warfighters or the civilian elements. IWARS already has a behavior engine that is used to control the agents actions (activities). The behavior engine uses "conditions" as a means to transition between different activities within the simulation. While the following may be a subtle distinction, it is important. There are cases, both in the simulation and with real SCU leaders, algorithms can be utilized to provide information that is useful to their selection of a course of action. |
Q: The SBIR description made reference to “100,000 iPhone applications.” |
A: The intended deliverables are not a set of smart phone applications. |
Q: Trained soldiers will respond differently to combat than an untrained mob. Suicidal fanatics will likewise respond differently. Also, some potentially dangerous situations can be defused by taking appropriate actions, while hostile encounters could erupt when inappropriate actions occur. To what degree are human/cultural behavior analogs expected to be fed into the developed algorithms? |
A: Human/cultural behavior analogs are expected to be fed into the developed algorithms only to the extent that they are important to the decision that is being supported. In many cases, human/cultural behavior analogs are not important information elements relating to the decision being supported. The following was excerpt from the topic, "There are common elements in the development of many simulation algorithms and those needed in decision support applications. In each case we would need to: identify the decision that we are trying to represent or support; identify the factors that are important; identify and develop methods of obtaining the needed data; develop user interfaces; develop methodologies and computer algorithms; and address platform integration issues." This highlights that we need to start with the decision that we are trying to support, then figuring out what the critical information requirements are and how we can work to provide them. Each decision will be supported by different information needs, which may or may not include human or cultural behavior analogs. |
Q: The topic of artificial intelligence is very broad (Genetic Algorithms, Neural Networks, Swarm Intelligence, Fuzzy Logic, Heuristic Learning, etc.). Is there any area of research that should be avoided due to integration issues with existing simulation environments or similar concerns? |
A: I don't think so. Please be advised that we are not trying to develop AI as part of this topic, but are trying to utilize algorithms to provide important information that can be used to support selection of a course of action, within the simulation or the SCU. |
Q: NOTE: Ref. 3 document has been uploaded in SITIS 8/18 and is now available for view/download. |
A: o |
Q: 1. Is it possible to obtain an evaluation copy of IWARS in advance of submitting a proposal? |
A: Unfortunately it is not possible to provide a copy of IWARS prior to the submission deadline. This topic has generated a great deal of interest and we are unable to provide copies in a timely manner to potential bidders. |
Q: Other than IWARS & OneSAF -- are there other special purpose constructive simulations that ought to be considered in planning the the Phase I research? |
A: The three main constructive simulations the Army uses for operational analysis are: IWARS, OneSAF and CombatXXI. IWARS is led by NSRDEC and AMSAA and focused on analysis at the small combat unit level. OneSAF is led by PM OneSAF and supports both training and analysis. CombatXXI is led by TRADOC Analysis Center (TRAC) and support their analysis mission. The Army uses and develops numerous other models and simulations. As a result, there may be other models or simulations that would benefit from this work and could be included as a target for this work. However, no other simulations are being explicitly identified by the SBIR topic author as an SBIR target at this time. |
Q: In the Phase I task solicition -- you ask for a solution that will provide " ...less involvement by the military analysts..." . |
A: The "analyst" term can be interpreted broadly as the best analyses are usually accomplished by multidisciplinary teams. It should be noted that IWARS, and potentially other target simulations, do not require the use of controllers. |
As of midnight September 1, questions for solicitations SBIR 10.3 and STTR 10.B will no longer be accepted.
To read the solicitation for full proposal preparation and submission details click here. |