Hi,
I'm looking for ideas re: inexpensive, close (near) proximity detection schemes. Ideally, without reliance on obvious "external" sensors (e.g., a video camera can be used but its presence is often very noticeable). And, small/cheap.
Things like capacitance (as used in the Theremin) have restrictions on *what* can be sensed. Ditto Ir schemes. E.g., how would you sense the presence of a bird in flight? A helium balloon? A ceramic globe? etc.
A pulsed IR laser diode and a tuned detector with a suitable optical filter?
How about ultrasonic transmitter/receiver and some cheap processing?
Hmmm... I hadn't considered "directional" techniques. I'd been thinking more along the lines of "how do I sense proximity to this *thing* (point)". Hence the Theremin came to mind.
I'll have to reconsider what I can/can't do in light of that! Thanks!
What kind of restrictions are you looking for, or are you thinking very broadly, fishing for options?
I think the best omnidirectional, close-in proximity will be something like a Theremin. It's not just for conductive things, since anything affecting the local electric field counts; hence, the "radar cross section" (if you will) of a metallic or conductive object will be large (actual size), while that of a dielectric (say, a solid nylon rod, or heavy PVC pipe) will be smaller, but still detectable. That of something with a very low average dielectric constant (foam, thin film?) will be very difficult or impossible to detect, even in direct contact.
Note that gradient matters, so if you put a hunk of dielectric in the close-in field, it will be partially shielded by the expanse of the antenna's field to begin with. Like how supermassive black holes have less shear at the event horizon than small ones (yeah, a familiar, intuitive analogy, I know..). A small antenna is desirable to sense smaller objects at closer distances. (This doesn't apply for large conductive objects, which short out all that field.)
If it's to detect something rigid, it's hard to go wrong with physical contact; it could be the deflection of a lever (e.g., cherry switch, "cat's whisker", ...), or the deflection of an air stream (pressure, direction, turbulence?), or... who knows. It has to apply some force, sooner or later, so could tear very thin film (or get broken itself), and probably will eventually abrade things (e.g., if it's sensing a large solid angle by spinning around quickly, you probably wouldn't want to put your eyeball up close to it..)
Reflection works for reflective objects, but you need a frequency where that's good; microwaves might offer reasonable performance without being especially directive (like optics) while being potentially more sensitive (i.e., having around a 1/4 wave range to pick up some real good con/destructive interference?). Same dielectric limitations would apply.
Doing it optically might be too complicated; simply monitoring diffuse light upon an illuminated sensor and detecting reflection or transmission changes would be subject to pretty strong limitations on the nature of the object (is it white or black, transparent or opaque, and how much of a feel do you get for its "cross section" as a result?), and line-of-sight (even with diffusers). You could get infinitely fancy, like using scanning lasers and such (which are indeed used for spacial 3D scanning heads), with some similar limitations, but by going into the serious power of radar/lidar stuff, that can overcome some of those limitations (e.g., if the return is weak, but we know its distance, that's still telling us exactly where it is; or if there's a hole in one spot, maybe it's reasonable to interpolate from nearby distances).
And if you want to get *really* esoteric, there are more types of radiation than electromagnetic and thermalized neutral baryon ensembles (i.e...poking it with matter). Neutrons are hardly biologically friendly, but have the potential to tell you a whole hell of a lot about the subject of interest (neutron activation and subsequent gamma and beta decays; possible neutron diffraction / scattering). Alphas are big burly ions that bounce around things; maybe useful in short distance (smoke detector?) or vacuum conditions.
As for inexpensive, these latter options certainly aren't going to fit, but it'll be damned hard to beat a ten cent webcam (with suitable software -- might not be inexpensive if NRE is required!), which can capture and
*identify* birds at much greater than "close in" distances. Quite a lot of effort has gone into hiding them; pinhole types are cheap and plentiful, and easy to disguise.Though there is something morbidly attractive about shooting high energy particle beams at flying things.
Tim
-- Seven Transistor Labs Electrical Engineering Consultation
Ultrasonics? Radar? IR proximity (gesture recognition sorts of sensors)?
Hack a shop bar code scanner, use an IR laser diode. Look at the analog return signal?
You should look at the auto industry, they're using pedestrian and collision detection systems based on camera, sonar, radar, lidar, time of flight. In the next few years these should become cheaper and better. Probably a combination of camera + another technology will be required.
The latter: "if I could sense THIS, then I could do THAT"...
The problem with capacitance is that you don't get a signal that is proportional to distance, etc. -- "closeness" -- but, rather, a "composite" signal. Like being only able to (visually) "see" large BLACK nylon rods ONLY when they are "very close" -- yet a person's *hand* at a much greater distance. :-/
It gets "nonintuitive" for people to try to understand the criteria governing the effectiveness of detection for different materials/objects. Refering to the black rod/hand above: they don't see any significant difference in how their (human) vision system -- which is how most people relate to the concept of "detecting proximity" -- between the black rod and the human hand. They can easily comment as to the relative locations of each without care for the *composition* of each.
[Note that I'm not trying to detect multiple objects but, rather, offering the two in this example as a way of illustrating how *people* "see" the two as 100% equivalent objects when it comes to sensing proximity.]That puts a limit on range. And, if you're trying to think omnidirectionally (for the time being), it gets more difficult not to end up with the sensor being very perceptible.
"Contact" can also suffer from its "ability" to detect "non things" -- e.g., is a jet of air impinging on a cat's whisker something you (I) want to be able to detect? Or, is it a false positive? (Ans: no, I hadn't planned on that)
And changes in ambient conditions, etc.
No, optical really only makes sense if you can "recognize" what is causing the change in the field. Compound that with cheap and unobtrusive and it's just a poor choice. What works weel for people doesn't necessarily work well for "machines".
Yes, but they still "look out" -- even with lenses to increase the field of view; or "look at" -- from afar (different perspective). Illumination then becomes an issue as well (even Ir illuminators can have some components that fall into the visible band)
It's easy to see how the Theremin idea (capacitance) can be so initially appealing (cost, unobtrusiveness, omnidirectional, etc.). Maybe I should hack something together and see just *how* limiting its technology ACTUALLY would be ("in practice"). Sometimes, "theory" just gets in the way...
Currently, there are *many* "other technologies" required -- not the least of which is the human component! :-/
It is always humbling to see how "complicated" seemingly "simple" problems can be!
Like, put a housefly on a leash, and if it tugs the leash, maybe it just saw something. You'll get a few false positives, of course... but it's dandy at detecting flyswatters.
Motion detection.
Hall effect? How close?
-- Les Cargill
Ah, so you're *not* just "fishing" -- you have some pretty strong assumptions in mind already. Namely, that it should be *visual* distance -- in which case, you'll have a hard time finding anything else but visual to prove the same fact. In that case, it will be all about detector design and illumination field.
What if you inflated a clear balloon/bubble and detected pressure changes? Near enough to invisible...
Another question: if people can't see it, is it still proximal? Example: very thin bubbles or foams, different materials under water (matched indices). (It's hard to find matching indices for solid materials in air, but the principle is the same.)
Sonar would be capable of detecting solid objects (perhaps a squishy balloon wouldn't be considered solid by some opinions?), without regard for the color or refraction of that object.
Although, that depends on sensitivity -- one might suppose that, whatever caused that air movement, probably has to be reasonably close by (e.g., a mouth blowing the air from only a little further away, or a compressed air nozzle from somewhat further still).
Tim
-- Seven Transistor Labs Electrical Engineering Consultation
40kHz Sonar (used for motion detection in alarm systems) is almost too sensitive... detects drapes moving from A/C turning on and off. ...Jim Thompson
-- | James E.Thompson | mens | | Analog Innovations | et |
No. Just that people need to be able to understand (in the sense of "grok") the "rules". E.g., PIr detectors seem magical to folks because the "rules" aren't intuitive (in the sense of the sorts of intuition that *they* would normally apply to explaining -- to themselves -- how something works or can be expected to BEHAVE)
Only in that people relate to "visual distance" more than they relate to "changes in the dielectric near the detector".
You relate to distance (proximity) of *audio* events by the intensity of the signal (AFTER you've characterized the source, in your mind). I.e., you'd imagine hearing words at a low volume (intensity) as indicating that the source (speaker) was nearby. You'd hardly assume it was a person speaking over a 5,000W PA system half a mile away (you would pick up other cues beyond just "intensity" to clue you in to that possibility).
You do similar processing visually. E.g., you'd not assume a "short person" was "farther away" (than a "normal height" person) *unless* you had no other cues to tell you that the person is, in fact, of a particular height. Of course, a short person standing beside a car that had been intentionally scaled down in size would confuse you (into believing both were farther away than they actually were). But, (particular models of) cars aren't available in 95%, 90%, 85%, etc. scales!
I thought about that. Even considered a clear/viscous fluid that I could sense motion through. But, then you're back with touch...
Ideally, a glass sphere would be detectable. E.g., if the glass sphere crashed
*into* the detector, few folks would claim it "wasn't there"!Yes, but sonar is rather directional. I suspect you could mount a transducer(s) and carefully characterize the return reflections, then look for differences. (i.e., so the detector would be as functional in "free space" as it would in an "enclosed space".
But, beyond "something has changed", I don't know how much information you could get from that approach (unless you could also characterize all of the "intruders"?)
I can't see how anything (other than stereoscopic vision, etc.) would be able to compensate for all the variations that could be encountered (e.g., big object at greater distance looks like smaller object at closer distance -- where "big" and "small" need not relate to actual mass/volume)
I'm going to have to think on this a bit more and see if I can narrow the constraints so I can exploit other, more "focused" technologies.
I'm looking for proximity -- so, change in position over time (i.e., motion) -- but the intervals are undefined. E.g., you could be looking at anything above "DC" :-/
I'm looking in the "foot or two" range (though that doesn't mean I don't want to be able to tell the difference between 2" and 4"; or 1' and 2')
I just had an idiot idea about sensing the gravitational force created by the object you are sensing.
How about we start from the beginning: active or passive? If active, then you emit "something" - probably EM waves - and look for the disturbances. If passive, you are either detecting something emitted by the object you are sensing (like infrared) or are detecting the disturbances caused in fields created by other sources - like a camera detecting bounced off light.
I don't think you can use a single method to detect something as wide as you are attempting to detect. A helium balloon (especially the see- through variant) and a ceramic globe are pretty different, not to mention a bird or a person.
Ah, so not presence or absence.
Yeah, that's a range that I have no background with.
-- Les Cargill
Correct. Sorry, I guess I hadn't considered that as a possibility.
Ideally, I want to be able to say "something" is XX units of length away from the target. I would settle for "something" is half as far away ("twice as close") as it was some time previous.
And, in all cases, would like to be able to either *know* that the item has come into direct contact with the target -- *or*, physically prevent this from happening.
Nor I. "Looking far" is a lot easier -- because you usually have more options at your disposal. Or, looking *really* close.
The Theremin was the only "thing" I could relate to that reacted to this sort of events.
I will hack together some kit and see just what sorts of materials, configurations, etc. "give it (capacitive) problems". Maybe I can just change the problem statement and sidestep the technical difficulties! :>
Theremins are very nonlinear.
This sounds like something where you're either doing interferometry or measuring something akin to Doppler shift. Sounds expen$ive :)
-- Les Cargill
Yes, but you don't care. As long as you have a mapping function (and repeatability) that can be consistently applied. The problem is how that mapping function varies from material to material, situation to situation, etc.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.