TECHNOLOGY AREA(S): Information Systems

ACQUISITION PROGRAM: Program Executive Office Integrated Warfare System (PEO IWS) 5.0 � Undersea Systems

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: Develop innovative big data analysis tools to detect, classify, and localize acoustic signals from torpedo-like targets.

DESCRIPTION: Moderate to high interference sources such as merchant ships and biologics often obscure threat signals evident on sonar display surfaces. In these situations, automated detection and classification of targets is especially challenging. However, an operator can typically identify potential threats amidst surrounding interferers if they focus on the bearing of the target. This potentially introduces unacceptable delays in operators� abilities to identify the bearing of that potential threat. The most time-critical threats are threats from torpedoes and rogue surface craft. The current level of automation predominantly trains on submarines and U.S. exercise torpedoes. Automation that detects torpedo-like threats needs to be optimized to remove delays in identifying these threats in moderate to high interference situations.

Deep learning, automated machine learning (ML), and big-data techniques have facilitated voice and facial recognition technology in devices such as cell phones, home security systems, and surveillance systems. The techniques used in these devices require massive amounts of data for training and testing an algorithm, the basis for computer metric analysis. While significant computational resources are required to process the data during the algorithm development phase, the resultant algorithm is fairly lightweight and portable. Modern sonar systems generate massive amounts of data. For example, the AN/SQQ-89 A(V)15 Undersea Warfare Combat System creates several hundred surfaces for automation and operator interrogation.

The Navy seeks development of innovative tools that provide timely and accurate detection, classification, and localization of threat targets; improves operator proficiency; and reduces the detect-to-engage (DTE) timeline. Through the use of big data analysis tools, the Navy seeks to expand current capability to better detect rest of world (ROW) threats and generically exploit passive acoustic characteristics present in all torpedo-like threats. An innovative approach is needed that will apply deep learning, ML, and big data techniques to acoustic and/or display-ready surface data to identify and localize threat targets in the data.

In Phase I, the developer will use representative, open source, Waveform Audio (WAV) files containing in-water interference sources (e.g. shipping noise, biologics, etc.) and a target of interest (e.g. high speed motor boat ) for which locations and identification of both interfering sources and the target of interest are known will be used to determine technology performance. The technology developed will be incorporated into the existing digital signal processing chain to support a high probability of correct classification and a low false alert rate to support existing operator displays. The technology should be capable of at least 70% correct classification with a false alert rate of no more than one (1) per hour in a semi-cluttered environment (e.g., a combatant in the presence of two surface vessels, two or more bathymetric features, and one target). Achieving a false alert rate of no more than one (1) per hour is especially important and will be a key metric in performance assessment. Transitions of these solutions to a tactical baseline will improve overall ship survivability in mission-critical situations. Initial technology transition is targeting the AN/SQQ-89A(V)15 Advanced Capability Builds (ACB) for U.S. combatants and other platforms performing Anti-Submarine Warfare (ASW) tasking. Therefore, it is important that the capability be feasible to integrate with a tactical sonar system. Open source WAV files will be used to evaluate the technology for SONAR application. Open source WAV files will not be provided by the government during Phase I but will be provided during Phase II.

The Phase II effort will 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: Define and develop a concept for detecting and classifying acoustic targets in interference using ML and big data analysis concepts. Demonstrate how the concept meets the requirements set forth in the Description. Establish feasibility through analytical modeling and simulation. Provide an estimate of the amount and type of data required to develop the concept for a sonar application. 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: Develop and deliver a prototype application that demonstrates accurate detections (within 30 seconds of initial target energy in water) of threat-like targets in semi-cluttered environments. Demonstrate, at a Government-provided facility, the prototype�s capability to meet the performance goals described in the Phase II SOW. Evaluate the prototype to show it is capable of processing single WAV files from classified tactical recordings containing interference sources and threat-like targets of interest as described in the Description. Classified acoustic recordings with associated truth information will be provided to the performer for prototype development.

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 technology to Navy use. It will be tested in an operationally relevant tactical baseline to determine if the tools meet the requirements of the AN/SQQ-89A(V)15 program in an integrated tactical system-level environment. Additional experimentation and refinement will be required during this phase. The prototype will be integrated into the AN/SQQ-89A(V)15 Program of Record. The product will be validated by the Test and Evaluation Support Group (TEASG).

This technology may be useful in commercial sonar applications such as marine mammal detection and tracking and underwater search and rescue applications.

REFERENCES:

1. Shamir, L. �Classification of large acoustic datasets using machine learning and crowdsourcing: application to whale calls.� Journal of the Acoustic Society of America, February 2014, 135(2): pp. 953-62. Doi 10.1121/1.4861348; https://www.ncbi.nlm.nih.gov/pubmed/25234903

2. Dia, Wei. �Acoustic Scene Recognition with Deep Learning.� Machine Learning Department, Carnegie Melon University. https://www.ml.cmu.edu/research/dap-papers/DAP_Dai_Wei.pdf

3. Halkais, Xanadu C. �Classification of mysticete sounds using machine learning techniques.� Acoustic Society of America, July 2013. https://www.ncbi.nlm.nih.gov/pubmed/24180760

4. Dugan, Peter J., Rice, Aaron A., and Urazghildiiev, Ildar R. �North Atlantic Right Whale acoustic signal processing: Part 1. Comparison of machine learning recognition algorithms.� 2010 IEEE Long Island Systems, Applications and Technology Conference (LISAT), 7 May 2010, pp. 1-6. https://ieeexplore.ieee.org/document/5478268/

KEYWORDS: Detection and Classification of Signals in Noise; Automated Machine Learning; Big-Data analytics; Deep Learning; Automated Detection and Classification of Torpedo-like Threats; Digital Signal Processing to Support Correct Classification