Triangulating Sound: Creating an Extensible Algorithm

[cnblock]

I would like to be able to triangulate a sound's location based on the inputs of two robotic sensors. I know this involves trigonometry, but I am a little out of practice. I think the practice of triangulation also is useful for radio signals. I would like to create a triangulation algorithm which can be extensible in nature, it just requires the addition of the amplitude, and it would in fact be encapsulated in an object as an algorithm.

 

[davenn]

hi you

I would like to create a triangulation algorithm which can be extensible in nature, it just requires the addition of the amplitude, and it would in fact be encapsulated in an object as an algorithm.

What research have you done for yourself so far ?
any particular links where there was something you didn't understand that you need help with ?
What level are you really at so people can give answers according to your level of understanding ? cheers
Dave

 

[skeptic2]

Tell us about your sensors. How would they determine the direction of the source accurately?

 

[Mark44]

I think the practice of triangulation also is useful for radio signals.

It is, but such a setup uses antennas that can be turned to find the direction in which the signal is strongest. For locating the source of a radio signal, one sensor detects the signal at a certain direction. Another sensor detects the radio signal in a different direction. By extending a line from each sensor in the direction toward the source, the point of intersection of the lines indicates where the radio signal is coming from.

I would like to create a triangulation algorithm which can be extensible in nature, it just requires the addition of the amplitude, and it would in fact be encapsulated in an object as an algorithm.

How you implement your algorithm (i.e., using object-oriented code or functional programming or whatever) doesn't have anything to do with your algorithm.

I don't see how getting the sound amplitude will be helpful. A very loud sound from far away would be as loud as one not as loud that is nearer. You wouldn't be able to gauge the distance based on the sound's amplitude. I once heard three booms that I thought were someone detonating dynamite nearby. It turned out that the sounds were caused by Mt. St. Helens exploding about 150 miles away.

 

[skeptic2]

It is, but such a setup uses antennas that can be turned to find the direction in which the signal is strongest.

The detection of a maximum is not easy. If the antenna pattern is broad is will be difficult to determine the angle of the maximum accurately. If it is narrow and the antenna is rotating, it will be difficult to prevent it from overshooting the maximum. It is much easier to find a null than a maximum. This can be done with two directional antennas pointing in slightly different directions, e.g., 30 deg. apart with their signals combined out of phase.

 

[russ_watters]

http://www.baltimoresun.com/news/maryland/crime/bs-md-ci-shotspotter-20180601-story.html

https://en.m.wikipedia.org/wiki/Gunfire_locator

 

[jim hardy]

Adding to Russ's list

thirty plus years ago i stumbled across an old old Army manual that described how they did it in WW1.

Wikipedia has a more modern description here
https://en.wikipedia.org/wiki/Artillery_sound_ranging

There's a more modern technique based on the concept of "Sound Intensity Vector" and i once saw such a machine.. It used Fourier analysis and clever signal processing math.
But that was before today's DSP IC chips , so it required a minicomputer that weighed the better part of a hundred pounds..
I think over the intervening years they've evolved "Acoustic Vector Sensors" that are hand held.
See if this article is a passable introduction to the technique
http://microflown-avisa.com/wp-content/uploads/2014/02/An-acoustic-vector-based-approach-to-locate-low-frequency-noise-sources-in-3D.pdf

I'm no expert, just was fascinated so remembered that it exists
Google some of the terms in that paper and see what turns up?

old jim