Department of the navy (don) 17. 1 Small Business Innovation Research (sbir) Proposal Submission Instructions introduction

TECHNOLOGY AREA(S): Air Platform, Battlespace, Sensors

ACQUISITION PROGRAM: PMA-263 Weather Hazard Avoidance, N2/6E Future Naval MetOc Capabilities

OBJECTIVE: Develop, demonstrate and transition a Navy Unmanned Aerial Vehicle (UAV)-compatible secondary payload for meteorological profiling on small and medium sized Unmanned Aerial Systems (UAS) to improve Electromagnetic Maneuver Warfare (EMW) and Intelligence, Surveillance, and Reconnaissance (ISR) sensor performance prediction, as well as aviation weather hazard sensing, avoidance and mitigation, and forecasting.

DESCRIPTION: Although small, inexpensive solid-state sensors have been developed for UAS applications, no meteorological payload exists commercially with adequate sensitivity, accuracy, and response time to measure the fine scale vertical gradients of absolute humidity, temperature, and pressure critical to assessing refractivity anomalies that affect electromagnetic propagation in the radar and communications bands. A payload with established performance capabilities and limitations is needed for multiple mission applications ranging in difficulty from on-board hazardous weather sense and avoid, to platform and propulsion performance, to assimilation into Numerical Weather Prediction (NWP) computer models, to ultimately a capability for single-station observed refractivity profiles equivalent to a research-grade rawinsonde. Challenges include design to fit small form factor and re-usability, a sensing and signal processing approach that accounts for flow distortion when integrated on a UAS that may not allow for the same constant ascent rate and consistent airflow that a balloonsonde or dropsonde has, and a design that allows for durability and re-use of most or all of the components as a UAS platform capability. By the end of the Phase II, sensor should be designed, integrated, tested and documented at the level of a Technical Data Package or Engineering Change Proposal for NAVAIR UAV Programs of Record to include sensor, housing, calibration and maintenance, and Interface Control Document (ICD). A potential commercialization strategy could be as an equivalent payload to the Aircraft Meteorological Data Relay (AMDAR) system, which currently provides profiles of temperature, pressure, humidity, and winds for commercial aircraft world-wide. A low SWAP, high accuracy integrated sensor package could add this capability to smaller aircraft and UAS and improve aviation weather analysis and prediction for safer aircraft operations. Additionally, with widespread use by the Navy, if the UAS Program Office picks this up for Phase III, would be a "tactical AMDAR" that could potentially greatly improve Carrier Strike Group level weather prediction in maritime operations where few other meteorological observations are available.

Nominal Performance Targets:

PHASE I: In Phase I a specific sensor engineering conceptual design and integration plan for Navy UAS (Puma, Scan Eagle, RQ-21) is required. Acquiring the UAS Interface Control Documents ICDs or technical specifications are the responsibility of the proposer and will not be provided by the government. Sensor component level and bread-board performance in a chamber is desired.

PHASE II: In Phase II, development of a brass-board sensor package in its near-final form factor and tested in a chamber with comparison to research quality reference sensors is required. A fully integrated sensor package flown on a UAS or manned aircraft proxy and compared to calibrated research quality reference sensors in a realistic environment is desired.

PHASE III DUAL USE APPLICATIONS: Final technical report will be provided to NAVAIR, Air Force, and NOAA UAS Programs of Record for weather hazard sense and avoid consideration as well as the Navy Information Warfare program for consideration of techniques to sense the electromagnetic spectrum environment. The expected Phase III transition is via a field change upgrade kit for Navy UAS Programs of Record and a COTS sensor available to other UAS applications in need of accurate, low-SWAP, high performance meteorological sensing such as the NOAA UAS Program Office, USCG Research and Development Center, and Department of Energy Atmospheric Measurement Program. Private Sector Commercial Potential: The Aircraft Meteorological Data Relay (AMDAR) system currently provides profiles of temperature, pressure, humidity, and winds for commercial aircraft world-wide. A low SWAP, high accuracy integrated sensor package could add this capability to smaller aircraft and UAS and improve aviation weather analysis and prediction for safer aircraft operations. Additionally, with widespread use by the Navy a "tactical AMDAR" could potentially greatly improve Carrier Strike Group level weather prediction.

REFERENCES:

1. The Aircraft Meteorological Data Relay (AMDAR) System. https://www.wmo.int/pages/prog/www/GOS/ABO/AMDAR/

2. Doyle J, Holt T, Flagg D, Tyndall D, Amerault C, Geiszler D, Haack T, Nachamkin J, Pauley P, Melville K, Lenain L, Reineman B, Statom N, Eber L, Roohi C. 2015. On the Impact of UAS Observations on High-Resolution Mesoscale Forecasts. 19th Conference on Integrated Observing and Assimilation Systems for the Atmosphere, Oceans, and Land Surface. 1.2. Phoenix, AZ, USA. 04-08 January 2015.

3. Flagg DD, Haack T, Doyle JD, Holt TR, Amerault CM, Geiszler D, Nachamkin J, Tyndall DP. 2015. The Impact of Assimilation of Unmanned Aerial System Observations on Numerical Weather Prediction Modeling of Modified Refractivity and Electromagnetic Propagation. (accepted) 2015 IEEE International Symposium on Antennas and Propagation and North American Radio Science Meeting. Vancouver, BC, Canada. 19-25 July 2015.

4. Guest PS. 2014. The use of kites, tethered balloons and miniature unmanned aerial vehicles for performing low level atmospheric measurements over water, land and sea ice surfaces. 18th Conference on Integrated Observing and Assimilation Systems for the Atmosphere, Oceans, and Land Surface. 8.6. Atlanta, GA, USA. 02-06 February 2014.

5. Mai T, Colon J, Molnar J. 2014. Virtual Field Test for Cognitive-Dynamic Spectrum Access Radios and Spectrum Usage Inventory. 2014 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting. 112.7. Memphis, TN, USA. 6-11 July 2014.

6. Pozderac JM, Johnson JT, Merrill CF. 2014. Measurement of S-, C-, and X-Band Propagation in the Marine Atmospheric Boundary Layer through Observations of Transmitters of Opportunity. USNC-URSI National Radio Science Meeting. Boulder, CO, USA. 8-11 January 2014.

7. Tyndall DP, Doyle JD, Holt T, Amerault CM, Flagg DD, Haack T, Nachamkin JE. 2015. Assimilation of UAS Observations from Trident Warrior 2013 into COAMPS NAVDAS. 19th Conference in Integrated Observing and Assimilation Systems for the Atmosphere, Oceans, and Land Surface. 10.3. Phoenix, AZ, USA. 04-08 January 2015.

8. Yardim C, Rogers LT, Gerstoft P. 2014. Verification of Trident Warrior 2013: Radiosonde and Numerical Weather Prediction Results with Passive Low Frequency RF Measurements. IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting. 502.9. Memphis, TN, USA. 6-11 July 2014.-

KEYWORDS: Meteorology; unmanned aircraft; radio propagation; maritime sensing; weather; electromagnetic maneuver warfare

Questions may also be submitted through DoD SBIR/STTR SITIS website.

N171-083

TITLE: Late-Stage Software Feature Reduction Tool for Security and Performance

TECHNOLOGY AREA(S): Information Systems

ACQUISITION PROGRAM: Total Platform Cyber Protection

OBJECTIVE: Investigate, design, and develop an automated or semi-automated software tool for the discovery, detection, and removal of unwanted or unnecessary software program features in any commonly used programming language.

DESCRIPTION: Modern commercial software is notoriously bloated due to the one-size-fits-all methodology commonly practiced in virtually all development and deployment efforts. This practice eases the burden on developers that intend to sell and deploy code to a large and diverse user base, but prior work has shown it can have a detrimental impact on performance and security [1, 2]. Many features built into a software program may not be needed by the average user, but are often included with no way for those users to disable or remove those features. Between the additional code (which may contain its own bugs and vulnerabilities) and the potentially undesirable functionality, extraneous features unnecessarily hamper performance while broadening a software product’s attack surface [2].

This effort seeks to reverse the trend toward one-size-fits-all software by creating prototype software tools that enable and empower end users to selectively remove software features they do not use or want. Examples of features to be removed could include elements of the user interface, support for legacy protocols, use of a camera or microphone, or something that could potentially compromise privacy such as a callback or diagnostic reporting functions. Some features may manifest themselves at the system call level while others may be more difficult to identify and trace back to specific regions of code. We make no assumption that developers have tagged their software to identify features, so identification of features and their corresponding code will be a key (but not insurmountable [3]) challenge of this effort, requiring performers to develop creative or innovative approaches in order to address it.

Current state-of-the-art efforts in software reduction have largely focused on methods to improve performance that do not modify the functionality of the original code [2, 4]. Performers of this effort must advance beyond the state-of-the-art to address removal of unwanted features, thus tailoring functionality to the end user. No current capability exists to selectively trim unwanted features from commodity software by automated or semi-automated means.

Given the focus on tailoring to the end user, tools proposed under this effort should be able to operate on software configurations commonly seen at delivery to the customer (e.g., APK files for Android or binaries for C/C++). The effort is not restricted to a specific programming language; submitters may choose to focus on any language for which they have the expertise so long as the language is general-purpose and commonly used. Both interpreted and compiled languages are of interest, but proposals should select a single language on which to focus.

The ultimate goal of this “late-stage customization” effort is to allow each end user to better customize apps and other software for their needs specifically and reduce both the bloat and attack surface of the software they run. The end users that will need to operate the tools can be assumed to be Power Users, but will not be program analysis experts. After identifying which features and corresponding code must be cut, the application in question must be transparently rewritten to produce a version with selected features removed. This process should occur in as automated a fashion as possible.

PHASE I: Develop a concept and methodology to identify features in software that may be of interest to remove and then tie them to their corresponding code. Develop a limited proof-of-concept prototype to demonstrate the viability of the approach for identifying and trimming software features.

PHASE II: Develop the prototype into a fully functioning software toolset for identifying and tagging features in general software applications of the chosen language, allowing end users to selectively remove unwanted features and their corresponding code. Demonstrate and evaluate the efficacy of the tools on several software applications of varying complexity as selected by the performer, along with demonstration of the continued correct and functional operation of the remaining features.

PHASE III DUAL USE APPLICATIONS: All third-party or commercial software used by the military contains extraneous features that unnecessarily widen a system’s attack surface. Being able to remove those features without needing the cooperation of the developer would be a great advantage and drastically help improve the security posture of such systems. As a result, expected transition of these tools could extend to a wide range of government programs interested in improving the security and performance parameters of their software environments. Based on the performer’s selected language, the performer will work with the Program Office to integrate their tool to the appropriate POR as the first transition target, which would be selected from C4I, combat, or control systems programs. Private Sector Commercial Potential: As cyber security concerns increase, end users (especially those in enterprise settings) will look to implement a more fine-grained reduction of unused or easily misused features in the commodity software they run. Also, with the decreasing rate of gains in hardware performance each year, users will also find value in tools that make their software more efficient and reduce their current computational and memory requirements. If successful, the solicited methodology and toolset would find great interested and a sizable market in the commercial sector, where the tools could be offered as a service to customize commodity software for end users.

REFERENCES:

1. N. Mitchell, G. Sevitsky, and H. Srinivasan. “The Diary of a Datum: An Approach to Modeling Runtime Complexity in Framework-Based Applications." In Proceedings of the Workshop on Library-Centric Software Design (LCSD), 2005.

2. Y. Jiang, D. Wu, and P. Liu. “JRed: Program Customization and Bloatware Mitigation Based on Static Analysis.” In Proceedings of the 40th IEEE Computer Society International Conference on Computers, Software & Applications (COMPSAC), Atlanta, GA, 2016.

3. Y. Jiang, C. Zhang, D. Wu, and P. Liu. “Feature-based Software Customization: Preliminary Analysis, Formalization, and Methods.” In Proceedings of the 17th IEEE High Assurance Systems Engineering Symposium (HASE), Orlando, FL, 2016.

4. T. Kalibera, P. Maj, F. Morandat, and J. Vitek. “A Fast Abstract Syntax Tree Interpreter for R.” In Proceedings of the International Conference on Virtual Execution Environments, New York, NY, 2014.-

KEYWORDS: software; feature reduction; Java; JavaScript; Python; C/C++; programming; efficiency; cyber; security; performance

Questions may also be submitted through DoD SBIR/STTR SITIS website.

N171-084

TITLE: Artificial Intelligence for Infantry Simulation in Small Unit Decision Making Training

TECHNOLOGY AREA(S): Human Systems

ACQUISITION PROGRAM: CMP-FY15-01: Accelerating Development of Small Unit Decision Makers

OBJECTIVE: Develop technologies to support the construction of Artificial Intelligence agents for use in simulations in infantry small unit decision making training. Agents must have realistic “thinking”, require minimal manual coding and editing of behaviors, and the tools must be capable of developing behaviors across a range of military infantry activities.

DESCRIPTION: Simulation is a key enabler for training that allows Warfighters to develop their skills without incurring costs associated with training (e.g. fuel, munitions, etc.) or putting their safety at risk [1]. The use of first-person simulations (e.g., Virtual Battle Space 3) for infantry training is manpower-intensive, requiring additional operators (i.e., pucksters) to control friendly and enemy military units, limiting the ability to training only individual unit leaders (e.g. squad) without additional manpower support.

Historically, the military has long supported the use of constructive agents and virtual humans [2, 3] –although their application within the Marine Corps has been limited [1]. Infantry small unit leaders need simulation-based training without the manpower costs associated with pucksters –such as training that leverages intelligent and behaviorally realistic agents that are easy to teach. Agents are autonomous entities capable of goal-directed activities based on their perception of the simulated environment, such that military personnel interacting with them would not require extensive training nor amount to a frustrating undertaking.

Advances in Artificial Intelligence (AI) have increased machines' capability to learn and allow machines to outperform humans on video games and board games [4, 5]. However, these techniques require a very large dataset for training and are mostly constrained to discrete domains. As such, these approaches are not easily extensible to the infantry simulation where rules and environments are ill-defined and action spaces are continuous. Aside from deep reinforcement learning, other approaches that complement traditional human learning have been investigated. For example, Interactive Task Learning (ITL) is an approach to support artificial agents learning new tasks through natural interactions and observations with humans [6]. The Air Force Research Laboratory (AFRL) has been developing a Synthetic Teammate capability [7], but there has been limited application to the infantry domain.

The goal of this effort is to develop technologies to support the construction of intelligent artificial agents in order to reduce the manpower required to develop and oversee the execution of those agents within infantry training first-person simulations. The focus of the training is on developing AI-based behavior for friendly and enemy agents to support Marine Corps Forward Observers training. In addition to the technologies for agent learning, there is also a need to develop metrics and evaluations that can span across multiple Infantry tasks and domains.

All development and demonstrations should be done with simulation engines that have no or minimal licensing fees for development or run-time execution (e.g. Unity).

PHASE I: Required Phase I deliverables will include a feasibility study. Included in this study will be an initial concept design for Artificial Intelligence for Infantry Simulation in Small Unit Decision Making Training that models key elements as well as a detailed outline of success criteria. Additionally, at least one behavior using the technologies proposed should be developed. Since access to Marine Corps personnel will not be supported during Phase I, surrogate tasks are acceptable for proof of concept. A final report will be generated, including system performance metrics and plans for Phase II. Ensuring an “open architecture” to allow integration with other military relevant systems (e.g. Augmented Immersive Team Trainer, Virtual Battle Space 3) will be considered a critical performance metric. Phase II plans should include key component technological milestones and plans for at least two demonstrations.

PHASE II: Phase II will include further behavioral development and evaluations with at least two evaluations from SME representatives that will be identified from the government. Required Phase II deliverables will include the construction, demonstration, and validation of a prototype system based on results from Phase I. All appropriate engineering testing will be performed and a critical design review will be performed to finalize the design and technologies before the evaluations. Additional deliverables will include: 1) a working prototype, 2) any associated drawings and specification for its construction, and 3) test data on its performance, in accordance with the demonstration success criteria developed in Phase I.

PHASE III DUAL USE APPLICATIONS: The performer will be expected to support the Marine Corps in transitioning the software products that enable the construction of intelligent and behavioral realistic agents for Infantry small unit training. The software products are expected to support and/or be integrated with existing Marine Corps training simulations (e.g. Augmented Immersive Team Trainer, Virtual Battle Space 3). Phase III tasks will include certifying and qualifying the system for Marine Corps use, delivering a Marine Corps user manual for the product, and providing Marine Corps system specification materials. Private Sector Commercial Potential: It is anticipated this technology will have broad applications in military as well as commercial settings. The use of artificial intelligence (AI) is continuing to grow, but it is currently limited to certain sets of tasks. For example, virtual reality is currently being used by some professional sports teams (e.g. NFL), but has limited or no AI application. It would be very beneficial to allow position specific (e.g. Quarterback) training that supports realistic agent behavior so that players can practice for upcoming games and specific opponents either with their team or independently.

REFERENCES:

1. Naval Research Advisory Committee. (2009). Immersive Simulation for Marine Corps Small Unit Training. Retrieved 6 June 2016 from http://www.nrac.navy.mil/docs/2009_rpt_Immersive_Sim.pdf

2. Pew, R. W., & Mavor, A. S. (Eds.). (1998). Modeling human and organizational behavior: Application to military simulations. National Academies Press.

3. Zacharias, G. L., MacMillan, J., & Van Hemel, S. B. (Eds.). (2008). Behavioral modeling and simulation: From individuals to societies. National Academies Press.

4. Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., ... & Petersen, S. (2015). Human-level control through deep reinforcement learning. Nature, 518(7540), 529-533.

5. Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., Van Den Driessche, G., ... & Dieleman, S. (2016). Mastering the game of Go with deep neural networks and tree search. Nature, 529(7587), 484-489.

6. Laird, J. (2014). Report on the NSF-funded Workshop on Taskability (Interactive Task Learning).

7. Ball, J., Myers, C., Heiberg, A., Cooke, N. J., Matessa, M., Freiman, M., & Rodgers, S. (2010). The synthetic teammate project. Computational and Mathematical Organization Theory, 16(3), 271-299.-

KEYWORDS: Artificial Intelligence (AI); Simulation Training; Autonomous Agents; Human-Robot Interaction (HRI); Machine Learning; Constructive Agents; Virtual Humans

Questions may also be submitted through DoD SBIR/STTR SITIS website.

TECHNOLOGY AREA(S): Materials/Processes, Sensors, Weapons

ACQUISITION PROGRAM: PMS-405, Navy Directed Energy and Electric Weapons program office, Surface Navy Laser Weapon System (SNLWS) Program

OBJECTIVE: The objective of this topic is to develop, design, construct, and deliver a compact, transportable instrument that can characterize ultra-short pulsed laser pulses including all spatial, temporal, spectral, pulse energy, and phase properties on a pulse-by-pulse basis.