Neural network modeling of a dolphin’s sonar discrimination capabilities

Abstract

The capability of an echo-locating dolphin to discriminate differences in the wall thickness of cylinders was previously modeled by a counterpropagation neural network using only spectral information of the echoes [W. W. L. Au, J. Acoust. Soc. Am. 95, 2728–2735 (1994)]. In this study, both time and frequency information were used to model the dolphin discrimination capabilities. Echoes from the same cylinders were digitized using a broadband simulated dolphin sonar signal with the transducer mounted on the dolphin’s pen. The echoes were filtered by a bank of continuous constant-Q digital filters and the energy from each filter was computed in time increments of 1/bandwidth. Echo features of the standard and each comparison target were analyzed in pairs by both a counterpropagation and a backpropagation neural network. The backpropagation network performed better than the counterpropagation network, and the use of both time and frequency domain features resulted in better performance than if only time or frequency domain features were used. [Work supported by ONR, Grant No. ONR N00014-93-1378.]