Go Down

Topic: Presence detection (sub 1 meter) on a budget  (Read 599 times) previous topic - next topic

Stilren

Aug 27, 2016, 10:46 pm Last Edit: Aug 28, 2016, 11:18 am by Stilren
Hi! I am working on a personal project* where I will have a small device on a persons work desk to check is the person is there or not. These are the requirements:

  • I will have to build around 30 of these so the technique used can not be too expensive (around 5 dollar for this functionality is ok).
  • The people will be in an open office environment and I would like to keep false positives from co-workers passing by behind to a minimum.
  • If a person leaves the desk I would like to be able to tell after 30 seconds or less
  • Some instructions on where and how to place the device can be given (e.g. this side should face you) but the detection range can not be too narrow so that false negatives are given


What I have now is a PIR sensor like this one (HC-SR501: cheap from china) But the range is too great, adding additional layers of plastic infront of the sensor helps but it get unreliable.

I have been thinking of ultra-sound but I am not sure the detection width of the standard HC-SR04 (link to sensor) is wide enough.

I have also been looking at smarter IR sensors like the one used in this clip . I have not been able to iddentify the sensor used but maybe MLX90614 (link to sensor) would do the trick?

Anyway, I would very much like to discuss this with the bright people frequenting this forum if you have any input! And yes there has been much googling.

*Edit: The project will be used in a master thesis in ergonomics. The people in the study will be volunteers and we will follow ethical best practice.

lastchancename

#1
Aug 27, 2016, 11:03 pm Last Edit: Aug 27, 2016, 11:04 pm by lastchancename
I've done something similar in the past with visible light phototransistors.
Variations in ambient light or reflections from the victim kept retriggering the sensor until the person left, fell asleep or died.
My solution had a wide view, but could easily be 'choked down' with a tube of black heat shrink.
To set a reliable range of hysteresis, i used an 8-pin micro to average events over time to eliminate false triggering.  It worked great.  Cost less than US$3 each.
Experienced responders have a nose for laziness, (they were beginners once)... Sure, there are trolls, chest-beaters, and pretenders - but the help you'll get here is about as good as it gets - if you try to help youself!.

Grumpy_Mike

Quote
Anyway, I would very much like to discuss this with the bright people frequenting this forum
I suspect that the bright people of this forum would find your project distasteful.

larryd

Quote
personal project
Quote
build around 30
Doesn't sound like a personal project.

The things you refer to are reasonable, you need to experiment and do some fine tuning to get things to work.

Using a SR04 and coding things to have at least one measurement near by (example 3 feet) within a period of time should work.

Perhaps a warning beep prior to alarm, to warn people.

Note:
Suggested this in a similar situation once before, everyone evolved refused to be monitored as it was too much like big brother was watching.


.
No technical PMs.
The last thing you did is where you should start looking.

6v6gt


Stilren

I've done something similar in the past with visible light phototransistors.
Variations in ambient light or reflections from the victim kept retriggering the sensor until the person left, fell asleep or died.
This sound really interesting! I will PM you about more information. Thanks.

I suspect that the bright people of this forum would find your project distasteful.
I think those people should not be so quick to jump the gun there. The application of this project is in ergonomics, its for a master thesis, all participants are volunteers and we follow appropriate ethical practices.

... Using a SR04 and coding things to have at least one measurement near by (example 3 feet) within a period of time should work. ...
I will try this, I have limited time and resources for experimenting in the actual environment though. Thanks

Have a look here for a similar request and some imaginative solutions:
https://forum.arduino.cc/index.php?topic=380567.0
Yeah I read that but I didnt find any useful information.

How about this sensor: MLX90614 its i2c and it seems to be a similar sensor to whats being used in the Ben Heck TV Show. Does anyone have any hands on experience here?


PaulRB

#6
Aug 28, 2016, 10:40 am Last Edit: Aug 28, 2016, 10:42 am by PaulRB
You should be spending your time looking for a way to measure how productive the people are, not how much time they spend at their desks. That is bad management. They could be asleep at the desk!

allanhurst

You could always use a pressure switch on the seat.

eg tractors won't let the engine run without a weight there...

But I think it's a bad idea...

regards

Allan

Stilren

You should be spending your time looking for a way to measure how productive the people are, not how much time they spend at their desks. That is bad management. They could be asleep at the desk!
I will update OP to better explain the project purpose so that we may discuss the technical issues.

 

jrdoner

How about a thermistor embedded right under the fabric in the chair?  Room temperature probably won't ever exceed, say 76 F. in an office, and when somebody plunks down in their chair, the thermocouple will probably read at least 90 F after, say, 2 minutes.   Bring the thermistor wires to the back of the chair to hook to a Nano and a cheap little RF module.  That might get your cost down to $5.00.   Of course, be prepared for the smart ass who sets his warm coffee cup in the chair as he leaves the desk.   Maybe multiple thermistors -- one under each cheek.

Go Up