Advanced Algorithms for Total Ship Monitoring Improvements
Small Business Information
12450 Fair Lakes Circle, Suite 800, Fairfax, VA, 22033
Director of Contracts
Director of Contracts
AbstractThe Total Ship Monitoring System (TSMS) facilitates real-time assessment of acoustic stealth for all SSN and SSBN submarine classes. A highly portable software baseline has been established that facilitates affordable hardware technology refresh andevolving capability improvement.Advanced Algorithm Improvements Phase I results demonstrate significant improvements to data processing algorithms and operator displays that make TSMS more tactically useful and less operator intensive. Efforts focused on advancements in 1) improvedassessment of vulnerability to counter-detection, 2) improved estimation of aspect-dependent narrowband and broadband transfer function, 3) improved transient detection, 4) enhanced automation of operator tasks to locate and classify offending acousticsignals, and 5) integration of Hull Array sensors into the TSMS for improved identification and localization of on-board and off-board sounds.Phase II tasks will continue to address the topics of Phase I. Signal processing methods will be evaluated using recorded data collected from operational deployments. Improvements to the accuracy of transfer functions by use of new combinations ofexisting sensors, new sensors, and new sensor locations will be studied. Algorithms promising significant improvements in the tactical value of TSMS will be implemented into real-time software applications and prepared for transition to the productionTSMS baseline. The technology applied to the transient processing system is especially suited to signal processing tasks. It can be adapted to a variety of signal recognition tasks by selecting a suitable set of extracted features and training models to recognizespecific characteristics. As an example, the recognition system to be developed on Phase II is directly applicable to the SQQ-89A(V)15 TRAFS detection of threat targets. Using a high performance set of extracted features, the proposed method is expected toreduce false-alarm rate by specifically recognizing categories of non-threat contacts.
* information listed above is at the time of submission.