I have a project that is to be placed in a public area, and that is to react to the presence of people: someone walking by or just standing nearby. It should react to a distance of about 2 meters around it (or at least in a target area of 120-180 degrees).
Tried so far:
IR distance sensors. Works well to detect my presence but has a too narrow field of vision, 5 degrees or so at most. Would need heaps of such sensors to cover the target area.
HC-SR04 ultrasound sensor. It just won't detect me reliably (clothing is probably too absorbant to IR). The sensor works great for solid objects.
capacitive sensing. I've reached over 1 meter but it's too sensitive to the other electronics (just changing the colour of my WS2812B LEDs in reaction to me approaching throws off the reading big time).
Looked at PIR, great field of vision but the 10m range is way too much for this project. Bringing that down doesn't seem to be easy what I read online, without a word on reliability of thus attenuated sensors.
IR distance sensors. Works well to detect my presence but has a too narrow field of vision, 5 degrees or so at most. Would need heaps of such sensors to cover the target area.
pert:
Have you looked at the RCWL-0516 microwave radar sensor modules like this:
That's a really interesting one, but as usual such an ebay link raises more questions than it gives answers.
This page gives a little more information about it.
Input 4-28V, but there's also a connection marked 3V3 on the board suggesting it can also be powered with 3.3V and actually internally works at that voltage.
OUT: apparently a digital signal; highest value just over 3V suggesting indeed internally it operates at 3.3V.
Range can apparently be adjusted by soldering an SMD resistor on that board (size not given). Not exactly user friendly. A 1M resistor reduces it from 7m to 5m; smaller resistor for greater reduction? I'll have some experimenting to do I guess.
It also seems to be able to read through a plastic container, which would make hiding it from view really easy.
I've been thinking about something like this as well, but it makes it impossible to hide the sensor, which is pretty important in this project. Something protruding is acceptable, something moving around definitely not.
Besides, I'd probably go with the "woosh-woosh" effect of Kitt. Preferably with a sample of the sound effect.
My understanding is that the automatic doors you see at the big box stores use the same type of microwave radar system.
wvmarle:
That's a really interesting one, but as usual such an ebay link raises more questions than it gives answers.
Yeah, it's quite surprising that happyness8966 didn't provide detailed documentation for the product! I just grabbed the lowest price listing on eBay. I'm sure that to the seller this is just a widget that they likely don't know a thing about. Well, I don't expect a high level of support when I buy something for $0.75 USD with free shipping!
They actually are available for $0.50 with free shipping on Aliexpress, and I think the ones I bought were discounted even below that because I bought them during the 11/11 sale.
pert:
Yeah, it's quite surprising that happyness8966 didn't provide detailed documentation for the product!
I know, that never happens on e-bay and the like
Got all the information I need by now, a set of 10 of those are in my TMall shopping cart (going price is about USD 0.50 each indeed), will get them in a week or so. Hope it works as described. Description does tick all the boxes, and that at a very low price.
I use a TFMini scanning +/-5 degrees X/Y off 90 degrees to produce an image out to 600cm. One servo, Z, sweeps left to right, Z and the X/Y servos is swept +/-5 degrees to create a 3d image stored in ram. Each degree is scanned 12 times per degree.
I use 1 ESP32 to maintain the X/Y/Z platform, with 3 servos, and a MPU9250. The data collected is streamed across a 1Mb/sec CAN Bus to a ESP32 WROVER-B, that processes the data; one process being display onto a 128X128 OLED screen. At 400cm, I can 'see' a 2 inch pipe.
The ESP WROVER-B, also does a Adafruit GPS, a CAN Bus network (there is a 3rd ESP32 that I am still working on for unit mobility), and other environmental sensors.
The WROVER is, also, starting to process the 'image' with the navigation data from the GPS, to turn that data into mobility data. WIP.
One of those robot arms makes a great X/Y/Z platform and a place to mount the TFMini, the MPU9250, a ESP32, and a power supply unit to run that part of the show; of course I did not put it together as per instructions.
The PIR range can be reduced by simply pointing it at the ground. The 60-degree cone looking down from a lamp post will define a small area. Or inset it into the ground, but a speck of dirt on the lens will be a major problem. You can also experiment with the "corridor" version of the PIR and put a cardboard shield over it to limit its field of view.
The Panasonic "grid eye" sensor is the solution for stationary people, where a PIR won't work. Maybe it will work for you. Adafruit has a breakout.
My current main target is the rcwl-0516 sensor. It's motion detection of course which should be enough for this application. It reading through an enclosure is just perfect for the overall outlook.
