Motion Detector Unit For Loud Objects

Hi all,

Is there any algorithm around to recognize Doppler Effect even if no
exact frequency is known?

Noisy objects in motion do have specific frequency�s; those
frequency�s are however sometimes different from object to object.
Any loud object in motion should cause alarm.

Any ideas to implement via DSP�s?

With compliments, and best thanks to any answer

Marino

Marino wrote:
> Hi all,
> 
> Is there any algorithm around to recognize Doppler Effect even if no
> exact frequency is known?
> 
> Noisy objects in motion do have specific frequency�s; those
> frequency�s are however sometimes different from object to object.
> Any loud object in motion should cause alarm.
> 
> Any ideas to implement via DSP�s?
> 
> With compliments, and best thanks to any answer

There is no magic. There is no scale that can read "more" or "less" 
without a specification of more or less than what. An object at steady 
speed moving in a straight line not containing the observer. Will have a 
varying Doppler shift. An accurate measurement of the variation might 
reveal both the speed and the degree of skew IF the assumptions hold.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Hello Marino,

> 
> Is there any algorithm around to recognize Doppler Effect even if no
> exact frequency is known?
> 
> Noisy objects in motion do have specific frequency?s; those
> frequency?s are however sometimes different from object to object.
> Any loud object in motion should cause alarm.
> 

That's exactly what medical Doppler ultrasound does. Modern ones 
calculate the FFT and then look at the changes in that FFT to 
automatically calculate some parameters, like diastolic to systolic 
volume and so on. There should be plenty of literature out there. I just 
don't have any since I kind of grew up with that stuff.

Regards, Joerg

http://www.analogconsultants.com

I thought Doppler Ultrasound measures the Doppler shift on the
reflected ultrasonic signal, and since the signal was originally
transmitted by the sensor, it's exact original frequency is known.

Mark

Joerg wrote:
> Hello Marino,
> 
>>
>> Is there any algorithm around to recognize Doppler Effect even if no
>> exact frequency is known?
>>
>> Noisy objects in motion do have specific frequency?s; those
>> frequency?s are however sometimes different from object to object.
>> Any loud object in motion should cause alarm.
>>
> 
> That's exactly what medical Doppler ultrasound does. Modern ones 
> calculate the FFT and then look at the changes in that FFT to 
> automatically calculate some parameters, like diastolic to systolic 
> volume and so on. There should be plenty of literature out there. I just 
> don't have any since I kind of grew up with that stuff.

Blood flow is intermittent. That provides a baseline.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Mark wrote:
> I thought Doppler Ultrasound measures the Doppler shift on the
> reflected ultrasonic signal, and since the signal was originally
> transmitted by the sensor, it's exact original frequency is known.

Right on!

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Marino wrote:
> Hi all,
> 
> Is there any algorithm around to recognize Doppler Effect even if no
> exact frequency is known?

If the object is moving on a straight line, with respect to the obsever 
one sees a frequency shift at the time of CPA.
> 
> Noisy objects in motion do have specific frequency�s; those
> frequency�s are however sometimes different from object to object.
> Any loud object in motion should cause alarm.
> 
> Any ideas to implement via DSP�s?
> 
> With compliments, and best thanks to any answer
> 
> Marino
> 

Hello Mark, Hello Jerry,

>> I thought Doppler Ultrasound measures the Doppler shift on the
>> reflected ultrasonic signal, and since the signal was originally
>> transmitted by the sensor, it's exact original frequency is known.
> 

For some reason I can't see Mark's post. Happens at times on the SBC server.

Anyway, what I meant was that these system measure the 'change' in the 
FFT response. This FFT response is not known beforehand. This response 
depends on the patient, the blood vessel size, the amount of 
calcification etc. It'll be different from one patient to the next.

Yes, the carrier frequency of the burst is known but it doesn't play the 
major role in this calculation. On most older system there is no 
calculation after the FFT but the experienced cardiologist or even 
technician knows after a while what certain patterns mean.

Mark, in your case it should be doable as long as the object doesn't 
perform a straight approach or departure towards the sensor 
(microphone?). You'll always have a 'change' in Doppler frequency. Those 
of us who have served in the forces know that when something gets 
dropped from an aircraft your hear a whine. The exact frequency of that 
whine isn't known a priori because it depends on a lot of factors that 
the guys on the ground don't know. But we do register the change in 
pitch. If the pitch doesn't change that means to get out of there, and 
fast...

Regards, Joerg

http://www.analogconsultants.com

On Thu, 27 Oct 2005 21:44:22 -0400, Jerry Avins <jya@ieee.org> wrote:

>Marino wrote:
>> Hi all,
>> 
>> Is there any algorithm around to recognize Doppler Effect even if no
>> exact frequency is known?
>> 
>> Noisy objects in motion do have specific frequency?s; those
>> frequency?s are however sometimes different from object to object.
>> Any loud object in motion should cause alarm.
>> 
>> Any ideas to implement via DSP?s?
>> 
>> With compliments, and best thanks to any answer
>
>There is no magic. There is no scale that can read "more" or "less" 
>without a specification of more or less than what. An object at steady 
>speed moving in a straight line not containing the observer. Will have a 
>varying Doppler shift. An accurate measurement of the variation might 
>reveal both the speed and the degree of skew IF the assumptions hold.
>
>Jerry

Yup, that was my master's thesis.   Listen to the airplane a little
bit, sort out the doppler behavior and estimate the speed and range to
the airplane.   Worked pretty well.   I demo'd estimating the speed of
cars at the Indy 500 from sound clips from a wall-mounted camera, and
did pretty well with clips of propeller airplanes and such as well.
Jets are problematic, as are helicopters, at least for the way that I
did it.

But to address Marino's q's, it's not hard to just look at the basic
relationships for Doppler behavior and see how to extract the desired
information (and when).


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

> If the object is moving on a straight line, with respect to the
obsever 
> one sees a frequency shift at the time of CPA.
> 

Your sentence is very interesting. Can you explain it please?

Do you suggest?
- Recording of two audio bursts over two times
- Then looking for any up-chirp or down-chirp between the two bursts
Or is it something else?

Marino