Evidence for acoustic imaging capability in a bottlenose dolphin

A necessary condition for acoustic imaging (e.g., synthetic aperture sonar processing) is the capability to sum echo samples from the same point in the environment over different signal-echo pairs. To decide whether a dolphin has this capability, a limited number N of electronically simulated echoes with constant delay were transmitted back to an echolocating dolphin. The experiment was performed in San Diego Bay, which has a large indigenous population of snapping shrimp. The signal-to-noise ratio (SNR) of the simulated echoes was controlled by adding artificial Gaussian noise and by varying echo amplitude. If the dolphin is capable of SAS-like imaging, then the SNR required for detection should decrease as the number of available echoes N is increased. This phenomenon was indeed observed. The best receiver model for describing the dependence of SNR on N uses binary summation (an M-out-of-N detector). Binary summation is robust against the strong impulsive interference produced by snapping shrimp. The effect of binary summation on SAS-like processing is assessed by creating SAS images from binary-quantized data at the output of a broadband, dolphin-like sonar, and comparing these images to those obtained without binary quantization.