The objective of this effort is to develop a conceptual design for a broadband binaural biomimetic sonar system that will scan in three dimensions and approximate the scanning behavior and success rate of echolocating dolphins. Current sonar systems, including those based on dolphin echolocation characteristics, exploit a single-point receiver. Even when arrays of receivers are used in extant sonars, their function is to increase the beam forming precision or to steer the sonar beam. Dolphins, in contrast, appear to use their two ears somewhat separable. Binaural processing adds significant information about the structure of the target in both time and frequency-domain representations of the signal. Studies with speech signals indicate a 5dB performance gain when using two receivers, relative to a single receiver. Similar improvements can be expected in the proposed biomimeteic sonar. To test this model, we will construct a binaural data-acquisition system capable of precise positioning and able to move freely in three dimensions as dolphins do when searching for fish. We will collect echoes from mine-like targets placed on a relatively clutter free sandy ocean bottom, and buried under the substrate and we will compare the performance of this system with current biomimetic monaural systems.
Keywords: DOLPHINS, TURSIOPS TRUNCATUS
