Need Advice on proximity sensor and capacitive touch

Other Hardware Sensors
1

Hi everbody,

This is my first topic so I hope i'm doing everything all right! Please let me know if not!

I have been working on an interactive art project for a bit now, and I've been struggling with a couple points, despite numerous tries.

You can see the art piece here : Geneva Lux Festival

To make it simple, it's a giant monolith, made of 30 blocks of concrete of 50 cm by 50 cm, with led strips in between. The main idea was to have an interactive piece of art, where people could interact with the structure.

Two main interactions that we tried were the following:

  • distance captors: as somebody approaches from or leaves the structure, the LED would dim accordingly
  • touch: we wanted to have an interaction when somebody touches one of the concrete blocks

None of the two interactions have been working properly, and after numerous attempts in doing it by myself with the help of google, I decided to ask the community for help!

Distance Sensors

For measuring the distances, we have several constraints:

  • sensor needs to be waterproof, dustproof, and works below zero, as the structure is outside for two weeks in the middle of january
  • sensor needs to be able to detect up to 2.5-3m
  • sensor should be small in size, 2 cm diameters
  • sensor should be accurate at detecting humans and
  • sensor should have a good refresh rate (There are 6 sensors around the structure to detect 360° and we want several readings (at least 2) per second so the structure is reactive)

So far I have tried different sensors :

  • HC SR04: the good ol' HC SR04, not waterproof
  • JSN SR04: too many false readings
  • A02YYUW: too slow
  • MB1000 LV-MaxSonar-EZ0 (that we put in epoxy to be waterproof): best one so far, good results indoor, but unreliable outside (maybe because of the ground made of soil?)

I also use a median filter to eliminate bad readings, but it wasn't enough... Lots of problem seems to arise from using ultrasound to detect people

Would you have any ideas on good sensors for proximity that fill the previous requirements / specifications? Maybe Lidar like the TFMini plus, i don't know if anyone ever used one for people detection?

Capacitive touch

We wanted each one of the bloc to detect when it was touched. Following this instructables, we decided to mix aluminum and iron to the concrete when it was poured, with a copper wire running inside, so the concrete would conduct electricity and detect human touch.

We invented a mechanism so the blocks wouldn't be in direct contact with the structure, and each block would be "electrically independent" , if it makes sense.

It works well when each blocks is separated from the structure, and kinda works when the blocks are on the structure, but definitely doesn't work when the LED strips are on. We believe electromagnetic noise caused by the LED mess with the capacitive touch sensors.

We used different sensors :

  • MPR121
  • CAP1188
  • AT42QT (1010, 1011, 1070)
  • even tried the embedded capacitive touch fonction of ESP32

After reading different articles on this forum, I tried to create a low-pass filter, as described here or here, without any success.
We switched the power supply to an external stable 3.3V and no change...

I'm running out of ideas, and would greatly appreciate helps on this matter...
Somebody advise me to look into inductive sensor, on one side of the block, to detect a hand that would be close by. What do you think?

I'm open to any ideas, I have been working on this project for 2 years, and it's driving me crazy... It seems that all my attempts fail, so really, any opinion is welcome :slight_smile:

Thank you very much

2

24 or 60 GHz radar "automatic sliding door" sensors.
Check the minimum activation distance specs.

Alternatively there are a bunch of Arduino compatible radar modules of varying sensitivity and directionality. You'll need heaters to prevent ice buildup.

3

Have you looked at garage door sensors, they are reliable and have a nice range. The ones I use transmit in current mode. I use a 150 Ohm resistor to + to power it then measure the voltage drop across the resistor. Nice part they are not to expensive if not purchased for a specific door.

4

I'm not seeing one in your post. Have you tried an infrared distance sensor? I assume they don't work well in daylight but I've not tried. They're cheap so testing would not be a big hurdle.

Have you considered using a Raspberry Pi (or similar) to detect "radio approaching" as the human (mobile phone) sensor?

5

24 or 60 GHz radar "automatic sliding door" sensors.
Check the minimum activation distance specs.

Alternatively there are a bunch of Arduino compatible radar modules of varying sensitivity and directionality. You'll need heaters to prevent ice buildup.

We would like to be able to know the distance at which a persons stands, I believe the "automatic sliding door" sensor would be just a on / off unfortunately...

Have you looked at garage door sensors, they are reliable and have a nice range. The ones I use transmit in current mode. I use a 150 Ohm resistor to + to power it then measure the voltage drop across the resistor. Nice part they are not to expensive if not purchased for a specific door.

Same problem, would we able to know the distance at which a persons stands?

I'm not seeing one in your post. Have you tried an infrared distance sensor? I assume they don't work well in daylight but I've not tried. They're cheap so testing would not be a big hurdle.

We didn't proceed with infrared because it usually needs a "receptor" on the other side, and since the structure is outside, we can't install one.

Have you considered using a Raspberry Pi (or similar) to detect "radio approaching" as the human (mobile phone) sensor?

Not sure if that would work properly outdoors, plus with the structure acting as some sort of Faraday cage. Same issue of measuring the distance properly...

6

Google Search using the terms from my post clearly demonstrating your assumption is incorrect.

7

Sorry, I didn't mean to be rude, I should have explained more!

I explored different solutions with infrared, but couldn't find one that suited our application, because of the following:

  • detection up to 3m
  • small in size
  • return the actual distance (the one I found would return 0 or 1 and not the actual distance)

If you click on the link you can see the actual structure, and we drilled 1 inch holes in the bottom block so we could put the sensor, so the very large IR sensor couldn't fit...

The other option was to put a "receptor" on the other side, which we couldn't do.

The closest one I found that could be a viable option was this one by DFRobot but the max distance was 200cm

We might revert to this one, as IR might be more reliable..

What do you think? Do you believe it would be good at detecting humans without too much noise? Do you know a good IR device that would be roughly the same size and return the actual distance?

I greatly appreciate all your help!

8

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.