TECHNOLOGY AREA(S): Information Systems

ACQUISITION PROGRAM: PEO IWS 1.0, AEGIS Combat System Program Office

The technology within this topic is restricted under the International Traffic in Arms Regulation (ITAR), 22 CFR Parts 120-130, which controls the export and import of defense-related material and services, including export of sensitive technical data, or the Export Administration Regulation (EAR), 15 CFR Parts 730-774, which controls dual use items. Offerors must disclose any proposed use of foreign nationals (FNs), their country(ies) of origin, the type of visa or work permit possessed, and the statement of work (SOW) tasks intended for accomplishment by the FN(s) in accordance with section 3.5 of the Announcement. Offerors are advised foreign nationals proposed to perform on this topic may be restricted due to the technical data under US Export Control Laws.

OBJECTIVE: Provide an Artificial Intelligence (AI) capability for target identification and behavior-based predictive track vector generation combined with Real-time Modeling and Simulation (M&S) based combat system targeting and tracking management that fills gaps presented in a communication and/or sensor challenged environment.

DESCRIPTION: The growing proliferation of unmanned air and surface vehicles poses a potential tactical threat to our current naval platforms. This threat continues to grow as the technology needed to develop such vehicles in vast numbers becomes easier for both state and non-state actors. The subsequent increase in the number of potential air and surface targets tracked and engaged in a tactical environment, where both communications and sensor data acquisition may be hampered by enemy activity, is of critical concern. The development of a mechanism that will provide the AEGIS (and potentially the Future Surface Combatant (FSC)) combat system with real-time M&S-based tracking updates to �fill in the gaps� when operating in a communication and/or sensor challenged environment is needed.

Non real-time (RT) M&S has been in use for a number of years within the Department of Defense (DoD) community. To date, its use has been more or less constrained within the analytical community and used to develop tactical engagement models for validating combat and weapons systems design requirements, tactical and strategic engagement modeling, and so forth. Recent advances in both AI (e.g., Deep Learning techniques pioneered by Google �Tensorflow� library & framework) and high-speed parallel computing architectures (such as the Nvidia and AMD Graphical Processing Unit (GPU) subsystems) may now provide the ability to execute M&S algorithms in a real-time environment. The potential of melding real-time M&S algorithms with known target behavioral models utilizing newly developed AI algorithms and techniques could yield significant tactical advantages. Current combat system tracking management algorithms utilize a simple linear predictive model based on last known position and velocity vector to update track data in situations where real-time sensor data is unavailable due to sensor failure or active sensor and/or communications jamming.

The RT on-the-fly track M&S agent (hereafter referred to as the RTS Agent) proposed technology is intended to function within a future tactical environment that may contain a large number of battlespace entities, all operating in a communications- and/or sensor-challenged electromagnetic (EM) environment. In such an environment, both organic and non-organic sensor data updates may be intermittently delayed or completely unavailable for an indeterminate period. Providing a capability in the combat system to estimate track position, velocity, and so forth during these intermittent periods will provide the commander with real-time modeling-based information. Such information will be of significant benefit in determining the actual current state of the battlespace. Additionally, an enhanced track picture will help reduce decision-related stress and fatigue by reducing the operator�s need to ponder over each track to determine its status, thus allowing for a potential increase in the operator�s ability to handle extended duty time and an associated reduction in manning, potentially improving affordability. The Navy seeks an innovative RTS Agent capability within the AEGIS combat system. The technology will be arbitrarily scalable to an indefinite number of battlespace-tracked entities, enabled by an architectural framework that leverages multiple parallel processing (in both hardware and software) of simultaneous tracks. Hence the final capacity would be determined by the parallel processing capacity of the hardware available at implementation. It will be relatively self-contained such that it will require only software running within its current host combat systems suite and be integral within the AEGIS combat system to provide complete single-platform based capability and have minimal to no impact on the performance of the combat system. It will also provide a well-defined and documented Applications Program Interface (API) allowing it to be easily ported to other combat systems architectures (i.e., SSDS and the FSC combat system) currently in the planning stage.

The RTS Agent architecture will be capable of creating estimated track vector updates based on RT track simulation and AI techniques utilizing prior track behavior and other data. This data will be gleaned from a combination of prior track behavior, its AI-projected target, and the tracked entity�s known capability. The latter will be determined from its estimated entity ID, referenced against an entity ID/Capability database.

The RTS Agent architecture proposed for development will be capable of presenting probability metrics for each potential predicted track (or group of tracks) in real time to the operator, with the goal of providing a >50% improvement in probability of successful target engagement when compared to the performance of an operator track picture unassisted or unaugmented by the RTS Agent. The developed RTS Agent architecture will be capable of coordinating its simulated track data with the track data of other platforms and performing multi-platform AI-assisted simulated track de-confliction, when such data is available.

The developed RTS Agent architecture will be capable of re-establishing simulated track synchronization with real-world sensor derived track data when such data again becomes available to the combat system either on an intermittent (asynchronous) or continuous basis. Both the developed RTS Agent architecture and any associated AI algorithms should be well documented, and conform to open systems architectural principle and standards. Architectural implementation attributes should include scalability, support of a well-documented open-systems API to support future capability upgrades, and the ability to run within the computing resources available within the AEGIS combat systems BL9 environment.

The Phase II effort will likely require secure access, and NAVSEA will process the DD254 to support the contractor for personnel and facility certification for secure access. The Phase I effort will not require access to classified information. If need be, data of the same level of complexity as secured data will be provided to support Phase I work.

Work produced in Phase II may become classified. Note: The prospective contractor(s) must be U.S. Owned and Operated with no Foreign Influence as defined by DoD 5220.22-M, National Industrial Security Program Operating Manual, unless acceptable mitigating procedures can and have been be implemented and approved by the Defense Security Service (DSS). The selected contractor and/or subcontractor must be able to acquire and maintain a secret level facility and Personnel Security Clearances, in order to perform on advanced phases of this contract as set forth by DSS and NAVSEA in order to gain access to classified information pertaining to the national defense of the United States and its allies; this will be an inherent requirement. The selected company will be required to safeguard classified material IAW DoD 5220.22-M during the advance phases of this contract.

PHASE I: Design, develop, and deliver architecture for a RTS Agent. Demonstrate that the concept shows that its proposed architecture framework and conceptual model can feasibly meet the requirements and parameters set forth in the Description. Establish feasibility through a study and/or use of a simulation-based analysis. Develop a Phase II plan. The Phase I Option, if exercised, will include the initial design specifications and capabilities description to build a prototype solution in Phase II.

PHASE II: Design, develop, and deliver a prototype RTS Agent that will demonstrate the capability to perform all parameters in the Description. Perform the demonstration at a Land Based Test Site (LBTS), provided by the Government, which represents an AEGIS BL9 or newer combat system environment and is capable of simultaneously simulating two AEGIS test platforms to allow for the demonstration of multi-platform simulated track de-confliction capabilities. Ensure that the prototype is capable of demonstrating its implementation and integration into the combat system environment. Prepare a Phase III development plan to transition the technology for Navy combat systems and potential commercial use.

It is probable that the work under this effort will be classified under Phase II (see Description section for details).

PHASE III DUAL USE APPLICATIONS: Support the Navy in transitioning the RTS Agent to Navy use as a fully functional software agent incorporated into the AEGIS combat system baseline modernization process. Integrate the RTS Agent into a baseline definition, validation testing, and combat system certification.

This capability has potential for dual-use capability within the commercial Air Traffic Control systems in situations when air traffic sensor data may be delayed or missing due to sensor or communications equipment failure.

REFERENCES:

1. Vasudevan, Vijay. �Tensorflow: A system for Large-Scale Machine Learning.� Usenix Association, USENIX OSDI 2016 Conference, 2 November 2016.� https://www.usenix.org/system/files/conference/osdi16/osdi16-abadi.pdf

2. Vasudevan, Vijay. �TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems.� Usenix Association, 2016.� http://download.tensorflow.org/paper/whitepaper2015.pdf

3. Schmidhuber, J�rgen. �Deep Learning in Neural Networks: An Overview.� Neural Networks Journal, Vol 61, January 2016.� http://www.sciencedirect.com/science/article/pii/S0893608014002135

4. Schmidt, Douglas. �A Naval Perspective on Open-Systems Architecture.�� Software Engineering Institute, Carnegie Mellon University, 27 March 2017.� https://insights.sei.cmu.edu/sei_blog/2016/07/a-naval-perspective-on-open-systems-architecture.html

5. Paquin, J. N. �The What, Where and Why of Real-Time Simulation.� IEEE, 3 April 2017. http://www.opal-rt.com/wp-content/themes/enfold-opal/pdf/L00161_0436.pdf

KEYWORDS: Real-time on-the-fly track Modeling and Simulation; Deep Learning techniques; Communications and/or sensor challenged electromagnetic (EM) environment; Artificial Intelligence; Multi-platform AI-assisted Simulated Track De-confliction; AI Agent-based Software Design