(Proposed project for the
University of Victoria)
Client: Microsec R&D Inc., Victoria
History: Microsec R&D Inc.
worked on this project for the Canadian Coast Guard about 10 years
ago. The idea was to identify the sound of humans
calling for help and to precisely locate that call using an array
of microphones with beamforming. A patented algorithm
that detects the human voice was applied to the beamformed
signals. The beam that did the best job gave the
exact direction and distance to the person.
There is also a wind protection system that may or may not be
available. The wind protection system is patented and
involves a rubber membrane stretched over the paraboloid
reflector. The wind striking the surface causes an
increase in pressure within the balloon which is constant for all
the microphones. The pressure on both sides of
the microphone membranes are balanced.
At the time the
project was done around 2000 the speed of computers was too slow
to do the job ... but only by about 15%. Computers
being much faster now are guaranteed to do the job.
Website: http://microsec.net
OBJECTIVE:
The objective of this project is to use an existing microphone
array to do the beamforming. analysis and display of the distance
and direction of someone calling for help. A
researcher in Vancouver Dr. Pierre Zakarauskas has designed a
voice detection algorithm that identifies human
sounds. It can be fooled by seagulls.
BACKGROUND: The microphone array that
has been constructed with a 2d parabolic reflector.
Beam-forming needs to be written in C (not
Matlab). The results of the beam-forming should
show graphically the direction and distance to a test
sound. Initially that test sound can simply be any
dominant sound of a reasonable intensity. Later the voice
detection software is added, allowing the system to automatically
show the distance and direction to the human identified sound.
Simple Early Array
Final Microphone Array Used for
Detecting People lost at Sea
Samples of
Array Processed Sounds
Using the Above Array (click them to listen)
Sample Sound before
Processing.
Sample Sound after
Processing.
SUBSYSTEMS:
1/
A regular computer is used.
2/
Amplification system using 8 op amps to amplify the channels.
This amplifier has already been designed but it could
be re-designed with newer hardware.
3/
Microphone array will be supplied.
CONSTRAINTS:
Since this is a rescue system, size, weight, and user
friendliness are important.
CREATIVITY AND INNOVATION: This project description is only intended as a starting point and we are open to new ideas. Moreover, there are still many unknowns that cannot be addressed until the project is underway. Microsec R&D Inc will be participating and conducting multiple experiments to provide insight into design choices and project direction.
BUDGET:
Microsec will supply any necessary materials, components, and
other resources (machine shop, experience, external resources or
consultants). To start off, we will provide a system
package. A budget of $1000 will be available for minor
electronic components.
-Graphical display of
distance and direction.
-Optional inclusion of voice detection algorithm.
-Software listings and
circuit drawings of any amplifiers or filters.
Contact Person
Ian
Soutar can be reached at 250 882 5896 (cell) or 250 853 3221
(Uvic).
email: soutar@uvic.ca
or isoutar@microsec.net