For my dissertation at UCSD, I’m working on a method of robotic sound source localization that uses multiple networked robots to graphically represent the physical and sonic characteristics of a given space. The current system uses a Beaglebone Black as its main processing engine.  For sound localization, audio is sent through an input board designed by myself and Mauricio de Oliveira.  The input board contains stereo preamps, signal summer/subtractor, envelope detectors, and threshold detectors. This allows communication with the BBB’s analog and digital GPIO pins.   So far, the robots have gone through several iterations, employing various means of control, audio input, and locomotion. Here are images of a few of those iterations.




20160420_104254 20160406_174607 20160316_223359

Screen Shot 2016-04-30 at 3.42.12 PM

This is the visual output of the graphing program. It provides a map of the robots’ movements and the physical and sonic environment,

Screen Shot 2016-02-12 at 10.51.23 PM

The audio input board layout

robot2 robot1

Screen Shot 2016-04-30 at 3.48.38 PM

A poster for a presentation at the Southern California Robotics Symposium

20160316_223312 20160316_223359 (1)

Robut5 Robut1 Robut2 Robut3 20160203_192429 20150103_214159 20141204_222231 20141204_222317