Acoustic Rapid Commercial Off-The-Shelf (COTS) Insertion (ARCI) Program

U.S. Navy selects MITRE

In April 1995, the Chief of Naval Operations Admiral Jeremy M. Boorda identified submarine acoustic superiority as the highest priority submarine modernization initiative. Following an internal Navy operational assessment, MITRE was asked to lead an independent Submarine Sonar Technology Panel to assess the technology development and transition process associated with submarine acoustic signal processing. In September 1995, the panel reported its findings and recommendations to the director of submarine warfare, Office of the Chief of Naval Operations. The primary conclusion was that the current passive sonar (i.e., sonar capable of receiving signals but not transmitting them) advanced development process was not effectively transitioning new technology to engineering development. It was determined that systemic process and product changes could enhance the current technological advantage the submarine force enjoyed over other submarine forces.

MITRE’s status as a Federally Funded Research and Development Center and our technical expertise and broad domain knowledge made our selection as the organization to work the passive sonar problem straightforward from the Navy’s perspective. The Navy viewed MITRE’s long-standing support to the service’s submarine communications and combat systems and our understanding of the technology being used in the submarine sonar community as advantageous.

While revamping the capability development process for submarine sonar, MITRE recommended the replacement of legacy sonar systems with COTS-based processors and implementation of a iterative, data driven process for incrementally fielding improved capability using advanced processing builds (APB). Advanced Processing Builds provide the technology transition mechanism to engineering development from exploratory development in government and industry science and technology activities. As the new process unfolded, MITRE conducted laboratory and analytical evaluations to identify specific technical approaches and algorithms that should be included in the first iterative cycle. In parallel, the Naval Undersea Warfare Center (NUWC) focused on fielding a series of near-term sonar augmentation packages to address critical operational issues. In November 1997, the first of four phases of ARCI installations occurred, and five months later the first APB was installed on an ARCI submarine for at-sea testing. Results from the entire program have been impressive.

Acoustic Rapid COTS Insertion

The ARCI baseline and a developmental APB are software programs targeted at a set of sonar arrays capable of running on the same hardware. The software exploits all opportunities to collect and archive submarine data. Selected data sets are used to benchmark ARCI performance and to measure the comparative performance implemented through an APB. The APB software is tested in a laboratory on real data as well as at-sea under controlled conditions. The most promising capabilities are deployed for operational testing, and those that add measurable value beyond the ARCI baseline benchmark are earmarked for fleet introduction. While data-driven successes in exploratory development will transition to an APB in advanced development, the APB itself will drive hardware technology insertion into ARCI if improved performance is required. Since both pieces of software use the same hardware, fleet implementation will occur at a faster pace.

One of the keys to the whole process is rigorous evaluation of candidate improvements using real at-sea data in the laboratory. MITRE led a groundbreaking effort to create the process and implement it in time to have a major impact on the content of the first APB. At the outset, there were a number of competing systems and capabilities being advocated by various pockets within the Navy sonar community. None of these systems had ever been independently evaluated against a common data set, while using consistent testing procedures. MITRE structured an objective and rigorous technical evaluation program whose results were ultimately briefed to Vice Admiral Richard W. Mies, Commander, Submarine Force Atlantic. Subsequent recommendations set the priorities for the APB program, and the capabilities that added the greatest value with little development impact were adopted for ARCI Phase 1.

In parallel with the laboratory evaluations, MITRE led an effort to create a physics-based predictive modeling capability at the algorithmic level. The work has developed mathematically-based models and analysis techniques that provide a better understanding of signal processing algorithm functionality, commonality, and performance relative to anticipated threat and noise environments. The intent is to baseline algorithm performance using analytically-derived performance bounds. The models provide insight into anomalies and uncertainties in system performance, guide the definition of performance metrics for test and evaluation, and help to ascertain the significance of system performance differences from testing using real data. The models reflect high-level statistical characterizations of the threat signals and noise environment, a statistical description of the individual algorithms within a processing chain, and performance measures such as Receiver Operating Characteristic (ROC) curves. Since the algorithms are not represented at a functional software level, real data is not used to drive the analyses. For general classes of threat and noise environments, the models are used to predict algorithm performance envelopes and to select processing chain configurations that can best exploit the unique capabilities of individual algorithms.

In addition to the national recognition attendant the Hammer Award, the team was further honored with a nomination for a MITRE Program Recognition Award. In his nomination of the team, Program Manager Ken McAdow wrote: "Project personnel performed the first-ever laboratory test of different sonar systems using a common set of sea test data. In addition and in spite of technical community skepticism, project personnel successfully developed detailed mathematical models of sonar systems. These models provided the sponsor unparalleled insight into algorithm performance. We have extended this effort to aid new algorithm developers in understanding the strengths and weaknesses of their approach as well as identifying alternatives. The team continues to evaluate each new sonar advanced processing build to help the government team produce a system of the highest caliber."

Page last updated: August 31, 1999   |   Top of page