robo-cop

Hello.

I am trying to build something like the "lego" robots out there in the Internet.
Well, at least my teacher prompt me to do. :roll:

I was looking for an analog schematic to localize a sound source in space (for robots), I came across with aspects such as "phase detectors" and "interaural time difference".

While in the Net, I found the pdf attached in this post.

My questions:

Can this be build out of discrete components?

How successfull could such effort be?

Any clues for a specific schematic?
You designers, have you ever came across such circuits?

In case it is impossible to design, would there likely be a chip out there that comprises that circuit?
To anybody knowing about such IC, I would like to know which is it, please.

To conclude, I believe that the circuit described in the pdf CAN be made by discrete components (in fact maybe not that difficult), my idea is:

Parts of this circuit remind me a lot the concept of PLL circuits.
So, I was thinking to integrate such PLL (phase-locked loop) chips, to use only certain components of them, and manage to build it.At least design it!

Can anyone here help me please?

Attached Files

analog_sound_localization.pdf (444.5 KB, 38 views)

Optikon

Well, the authors do a good job at describing the concept. I would be weary of the practicality of their design however. While they do a good job conceptually they only "simulate" their system with two input sine waves that are very clean. In the real world application, the situation will not be this nice & clean and the authors only mention the DSP that will have to take place to make a real system work. You will have to start wher ethe authors left off and it is probably the largest amount of effort yet to come.

This project could IMO take a year or more before a final real system is working with real applications.. not for the faint of heart! But you might be able to get some simple results in a controlled experiment just like the authors.

robo-cop

I know the whole process of such sound localization is time consuming.

I just want a schematic that can do in simulation what the pdf says.

I found intriguing the phase detector scheme with a very low "locking" time.

I want to perfrom a simple test with a simulation in the beginning, and then maybe deal with the subject some more.
So, clean sine waves are fine. I am for now exploring the concept of rapid locking time, as well as the 0 to 170 degrees locking range concept.
That's all.

Thank you for your time.

robo-cop

So, I am in need for suggestions.Like a circuit.Or suggestion for component selection. :roll:

When going through the pdf's circuit explanation, I saw that the circuit resembles very much the operation of a "phase-locked loop" (PLL) circuit.

Now, my question is, is there a way to inter-related some of the components of a PLL circuit (or chip) in order to achieve the "calculation of the incident angle" ?

I am assuming that this is a pretty straightforward question, I am eager to hear your comments. (I like brainstorming).

Optikon

Phase detector probably comes from a standard PLL circuit. You are looking to build this circuit? Each of the circuit blocks are standard circuits available on the web. You could start by putting them together.

The calculation of the incident angle is an arccosine relationship. The most efficient way to do this calculation in hardware is with a microcontroller or some other kind of processor. Probably the main DSP since you'll have on e of those in the real system.

I'm not sure where you are trying to go with more simulations. The authors already did the simple simulations, what will you be adding? And of what value will it be?

robo-cop

>>Phase detector probably comes from a standard PLL circuit. You are looking to build this circuit? Each of the circuit blocks are standard circuits available on the web. You could start by putting them together.

What I initialy intend to do is that.
I was looking here for (specific) suggestions for such a "puzzle".

>>The calculation of the incident angle is an arccosine relationship. The most efficient way to do this calculation in hardware is with a microcontroller or some other kind of processor. Probably the main DSP since you'll have on e of those in the real system.

>>Well, the authors think that an analog solution is better than the digital.
It is faster and has a wider locking range.
So do I.

>>I'm not sure where you are trying to go with more simulations. The authors already did the simple simulations, what will you be adding? And of what value will it be?[/quote]

All I want, as a first step, is to do the simulation myself in my simulation program.
So, I was hoping here somebody could assist me on that, by suggesting a complete schematic with component suggestion (values, parts).

Next, I could try to build it by integrating real PLL chip components;but to do that I need a complete comprehent simple schematic first. 8)

Thanks for the help in advance!