AAITS Teaches Undersea Acoustic Analysis to Navy Sonar Technicians
U. S. Navy
Sonar technicians and their instructors.
Traditional classroom instruction and computer-based training (CBT) are effective methods of presenting and testing undersea acoustic data analysis facts and concepts. However, carrying out acoustic analysis at expert levels of proficiency requires extensive practice analyzing many acoustic datasets, or LOFARGRAMs, under the guidance of experienced analysts. To train sonar technicians more rapidly and cost-effectively, the Navy needed advanced software which complements traditional training methods by providing a practice-based learning environment and automated instruction.
Stottler Henke developed for the U. S. Navy an Acoustic Analysis Intelligent Tutoring System (AAITS) which enables students to practice the detection and classification of sources of underwater acoustic signals. The system is composed of a Scenario Authoring Tool and a Tutoring System. Acoustic analysis subject matter experts create scenarios using the Scenario Authoring Tool by annotating LOFARGRAMs with significant features and links among related features and assigning a final classification. Students use the Tutoring System to request LOFARGRAMs, annotate them with features and links, propose final classifications, and request comparison of his or her annotations and classification with those of the expert. By comparing the details of each student’s analysis with those of the expert, the Tutoring System can identify the acoustic analysis principles understood and correctly applied by each student in order to provide specific and individualized feedback. By storing LOFARGRAMs annotated by experts, the Acoustic Analysis Intelligent Tutoring System also serves as a knowledge repository to disseminate the most current acoustic analysis expertise to sonar technicians on land or at sea.
AAITS has been deployed at ten sites within the Navy. The U.S. Navy designated AAITS as a Small Businesss Innovation Research Success Story.
The instructional method of automatically evaluating annotations of sonar data to assess students’ data analysis and interpretation skills can be applied to image, time-series, and other signal data. Related applications include medical image interpretation, visual defect inspection, and analysis of scientific, engineering, and business data.