Maker Pro
Maker Pro

Which type of sensor works for pigeons within 20 feet without tripping when a skytrain drives by

sundo597

Aug 21, 2020
1
Joined
Aug 21, 2020
Messages
1
at about 100 feet distance?

Hello, newbie here.
I want to get a sensor that will detect birds on my high-rise balcony per heading above so that the sensor will activate power to a fan that runs on regular 120 V AC power. The fan will then turn on and move tin foil to scare the birds away. Alternatively, I could use a battery operated fan if the setup is easier or the supplies are easier to obtain.

Thank You very much in advance for your help.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
PIR?
A111 Acconeer radar?
.
Mount it on the ceiling, pointing down at the floor?
 

ChosunOne

Jun 20, 2010
480
Joined
Jun 20, 2010
Messages
480
I very much doubt that what you want can be done practically.

Off-the-shelf motion sensors of the type used in alarm systems (or the less discriminating types used in turning on lights) can probably be used to detect pigeons at 20 feet out from your balcony. The problem is what's known in the alarm industry as the "elephant syndrome": A wasp flying an inch in front of a PIR [or microwave(MW) or ultrasonic(US)] sensor, creates as much of a "motion" signature as an elephant walking at 50 feet. (not an exact ratio, but you get the idea). Whichever of the three most popular motion sensing technologies you try to use is going to "see" a skytrain at 100' if it "sees" pigeons at 20'.

I put quotes around "motion" and "see" because "motion" sensors don't really sense motion. They sense the effects of objects moving in the space being "watched" by the sensors. For the PIR sensors, the effect is change over a short time in the heat signature of the background that the PIR is watching. That's why a PIR needs a minute or two after power-up to stabilize, to establish a baseline of what's "normal" for the background. For microwave and ultrasonic, the effect is causing a Doppler effect in the return signal so the sensor sees a MW or US frequency shift in the echo of a signal.

In all those effects, there is no way (that I know) that something large moving farther away won't be detected as easily as something small moving closer. Most models have a "range" adjustment that gives you the impression that you can set it to detect motion out to, say, 20''; but like the term "Motion Sensor", the term "range adjustment" is misleading: You can adjust a sensor to detect reliably a medium-size human (or pigeon) at 20', moving at ordinary speed (walking or flying), in a certain direction relative to the sensor, but factors like size, speed, and direction will change the range. It's really more of a "sensitivity" setting, but the point is that you can't set a "range" to have the sensor not detect the elephant/skytrain farther out, if you want it to see the pigeons at 20' out.

If you want to make it work at detecting pigeons much closer (not so sensitive)--like when they land on the balcony or railing or whatever, you might be able to manage it; especially if you're willing to aim the sensor on the balcony space instead of out away from the balcony. However, there are things you should be aware of. The "motion sensors" most probably _will_ be triggered by things other than pigeons from time to time. I know you have bats in Vancouver, and you can expect them to trip any motion sensor from time to time. Chance are you aren't even aware of bats visiting (most homeowners aren't aware of the bats that live in their attics), but they probably do. Also other passing birds, even large insects like moths & butterflies. (I mention this because of years of experience in chasing causes of false alarms in a closed and controlled environment!). A lot of the time, you probably won't even know what triggered your foil moving, because by the time you look, the "trigger" will be gone. So don't expect only pigeons to trip your assembly.
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
You could use e.g. a raspberry Pi with camera and computer vision trained to detect pigeons.
Not an easy venture, but everything you'll need is available off the shelf.
Some programming required, though.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
The Acconeer radar unit is at 60GHz. It will only propagate so far in air. This somewhat aleviates the distant elephant issue.
Doppler radar may be possible if the train you mention has a different heading to the pigeons and you can get the unit in line with the direction of travel.
Again, the geometry will likely matter. If there is an orientation that points workably towards pigeon arrivals and minimally towards trains, you may have something viable.
As mentioned
Some programming required, though.
Some assembly required :)
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
The most obvious solution: mount the sensor such that it can "see" the birds but not the skytrain. Find a different spot and correct the angle.
 

ChosunOne

Jun 20, 2010
480
Joined
Jun 20, 2010
Messages
480
The Acconeer radar unit is at 60GHz. It will only propagate so far in air. This somewhat aleviates the distant elephant issue.

Um, no. At whatever the frequency the radar/microwave (which is the same thing) operates, it's still subject to the inverse square law of attenuation that we all learned (hopefully) in high school: All other factors equal, a given reflecting surface at twice the distance will bounce back one-forth the signal; in this case, an object at 5 times the distance will reflect back one twenty-fifth the signal, for a given area of reflecting surface. So the more distant object would have to have twenty-five times the reflecting surface to return the same signal strength as the nearer one.

I think it's safe to say the skytrain shows more that twenty-five times the reflecting area of a pigeon--all other factors equal, which they aren't. I didn't bother to point out earlier that metal reflects radar better than pigeon feathers and will return a stronger signal. The train will probably not be coming/going straight toward/from the balcony (unless OP is located right at a curve in the track), and that oblique vector will lessen the Doppler effect; but at the same time, the train probably goes faster than pigeons flying full-out. I don't have the info I'd need to calculate the toward/away vector, but I'm still skeptical that any motion sensor can be used if it's aimed out away from the balcony. I could be wrong, and if you decided to try a standard "motion sensor", I'd like to hear how it works.

Doppler radar may be possible if the train you mention has a different heading to the pigeons and you can get the unit in line with the direction of travel.
Again, the geometry will likely matter. If there is an orientation that points workably towards pigeon arrivals and minimally towards trains, you may have something viable.

Yeah, good luck in finding a particular heading that pigeons will consistantly and reliably use. I'd expect them to come in and go out in all different directions, and at different speeds too. But again, I could be wrong. If you're going to use any kind of motion sensor, I think you're going to be constrained to aim it at the balcony, or perhaps the outer edge and obliquely out so that it doesn't "see" the passing skytrain.

If you decide to try one of the standard "motion sensors" you'll probably have a choice between microwave/radar and PIR. (I don't know if anybody uses ultrasonic motion sensors any more. I haven't seen it used in the alarm industry since the late 80s/early 90s) Each has its pros and cons/hazards that you should be aware of.
 
Last edited:

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
Um, no. At whatever the frequency the radar/microwave (which is the same thing) operates, it's still subject to the inverse square law of attenuation that we all learned (hopefully) in high school:

There is an absorption line in air at 60.5GHz. The attenuation increases dramatically about this frequency.
.
You message is a good one: You can't get over the large distance object versus the near smaller object, but the move of some things to 60GHz to contain the field of effect is also real.

Yeah, good luck in finding a particular heading that pigeons will consistantly and reliably use. I
A good point. I meant of course that the trains might have a consistent heading which might help to pay attention to. Again, not a total solution, but as with the general pointing direction, a factor to consider.

I think you're going to be constrained to aim it at the balcony, or perhaps the outer edge and obliquely out so that it doesn't "see" the passing skytrain.
Perhaps you put this better than I did. Thanks.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
Gee, folks... this is an application that requires RADAR! Distance-to-target detection is a fundamental property of RAdio Detection And Ranging or RADAR. There are several techniques used to discriminate among targets at different distances, but the earliest and simplest one was to measure the time required for a radio-frequency pulse to propagate from the transmitting antenna to the target and then return to the receiving antenna. The "rule of thumb" for this two-way propagation time is a "nanosecond per foot." So, if you "need to detect pigeons at 20 feet out from your balcony" while rejecting reflections from a skytrain at, say, 100 feet away, you will need very short pulses (on the order of a few nanoseconds) and a "range gate" circuit that can discriminate (and reject) signal reflections that come from targets more than 20 feet away. This is a strictly analog approach, but there are digital signal processing methods that would work too... just not exactly affordable at this time. The analog approach is not simple either, but it is doable. You could also frequency modulate (chirp) the transmit signal and use an inverse transform on the received signal (usually with a solid-state surface acoustical wave filter) to simulate a very narrow transmit pulse. And finally, there is Doppler discrimination that could possibly be applied to the frequency shift of the received signal, caused by the bird's flapping wings, to distinguish the birds from the background. This is just a little bit tougher, but the principle is the same as that used to implement Gunn diode "radar speed guns."

Lacking the wherewithal for any of the above techniques, a CMOS camera with IR LED illumination covering a wide field of view (FOV) could probably be used to detect birds. Aperture and focal length of the camera lens can be used to limit the depth of field (DOF) to blur anything further away than, say, 20 feet. Image processing software using unsharp masking and/or other techniques can then provide a go/no-go bird detection signal to activate a noise maker of some sort. I recommend something that provides a loud report and sprays some sort of substance towards the bird invasion, water perhaps. Wash, rinse, and repeat until the birds go away. Of course a human observer can be used to identify the bird invasion and trigger the noise maker. Perhaps after a few such incidents the birds will decide to roost somewhere less noisy or more amenable to their company. Or perhaps you could purchase or borrow a cat.
 

Nanren888

Nov 8, 2015
622
Joined
Nov 8, 2015
Messages
622
If you use FMCW then I guess there is pulse compression to a narow pulse (time-width about one over the bandwidth) or multiplication with the local reference to get a frequency that depends on range (assuming low Doppler).
The Acconeer radar I pointed you to, is interfaced as a digital device, has built-in detectors for peak-distance that tell you the range of targets & strength. You can set thw distance range to sweep & it will only tell you of targets within these range limits. That is, it does all the RF stuff for you.
I guess the NXP & TI parts that are similar probably act in a similar manner. Can't say I've used them.
 

hevans1944

Hop - AC8NS
Jun 21, 2012
4,878
Joined
Jun 21, 2012
Messages
4,878
Here is a brief blurb describing the Acconeer (http://www.acconeer.com) modular radar that @Nanren888 suggested. Picosecond (one thousand times better than nanosecond!) time resolution. Millimeter accuracy at ranges up to seven meters (twenty-something feet for those living here in the States) should easily detect pigeons while not detecting skytrains one hundred feet away. Best news: these puppies are affordable at about $13 each in small quantities.

Google for some hacker articles that describe how to use these fantastic little modular radar units. Much less fuss than a cat because no litter box and no cat food (in between pigeon snacks) is required. Thank you and kudos to @Nanren888 for bringing this new technology to our attention here on Electronics Point.
 
Last edited:
Top