EE Seminar: Dynamic Beamforming of Echolocating Bats

~~
Speaker: Pavel Kounitsky
M.Sc. student under the supervision of Prof. Anthony J. Weiss and Dr. Yossi Yovel

Wednesday, March 25, 2015  at  15:30
Room 011, Kitot Bldg., Faculty of Engineering

Dynamic Beamforming of Echolocating Bats
Abstract
This work focuses on the engineering aspects of the development of methodologies and analysis tools for multidisciplinary research on ultrasonic bats’ beamforming. 

Echolocating bats perceive their environment acoustically, by emitting ultrasonic pulses and analyzing the received echoes.  The volume of space that is covered by the ultrasound pulse and then is sensed by the bat depends on the emitted beam.  Echolocating bats can rapidly adjust many of their biosonar parameters to optimize sensory acquisition.  The ability to change the ultrasound beam in a functional way is an area of great interest in the fields of zoology, neuroscience and biomimicry.

The research of beamforming poses multiple engineering problems.  This study focuses on mouth emitting bats, that have been hypothesized to change the mouth gape to control the shape of the biosonar beam, and therefore, there are two main challenges: estimation of the mouth gape from a 2D camera image and beam shape restoration from an ultrasonic microphone array.

A neural network approach and a training algorithm have been chosen to estimate the mouth gape, solving both the camera’s angle and zoom variance problems.  Beam restoration involves application of multiple signal processing techniques, such as TDOA analysis, multilateration, flight trajectory Kalman filtering and beam interpolation.

With the developed set of tools, it was possible to perform correlation analysis over the observed features and make several dramatic conclusions about the dynamic beamforming of bats.