Self Organizing Machines - Applications to Underwater Rescue and Recovery,

Abstract : For the past several years Nortronics has supported an active internal research program in the area of self-organizing machines. As a part of this program, various applications to automatic pattern recognition have been examined, including machine classification of sonar signals. The importance of improving classification capability for deterrent/defensive naval forces has been recognized for some time. The development trend of sonar equipment has created a situation where the advantages of improved equipment performance (increased range and sensitivity, for example) has placed a corresponding burden upon the sonar operator's classification ability. Traditionally, the sonar classification task, regardless of the type of equipment; passive, active, processing, or display, always reverts to use of the human brain or a group of brains for final processing and classification. Therefore, substantial improvement in classification capability and speed can be expected where emphasis is placed upon those techniques which relieve the burden of classification on the human operator.