I want to detect if a seat is occupied by a person. This is the answer that I found on Stackoverflow.

However, how accurate is the output of 4 suggested sensors? Will the computing of all these 4 sensor values into one boolean recognize a case when, for instance, a box of 40kg is placed on a seat and some person is just slightly "shaking" that seat with a hand? In those cases Arduino should recognize that a seat is not occupied by a person. Will it work with only 4 sensors or is there any better and more accurate solution (maybe some existing library)? I guess that CapSense-Library is not applicable in my case.

  • All sensors have their downsides. And will sometimes falsely trigger. You can probably detect some false triggers with additional software. e.g. a load cell that continuously gives the same value would suggest a very inactive person, or a box. Also a box tends to not be not as heavy as a person (though maybe as heavy as a little child). In the original question cost was a major factor. In your case accuracy seems to be more important. Could you maybe elaborate on what it is you need to achieve? CapSense was actually be a suggestion that came to my mind when reading your title.
    – Gerben
    Oct 29, 2015 at 18:12
  • I guess that CapSense-Library is not applicable in my case. - why not? How about a heat sensor anyway? If the person is not wearing really thick clothing their rear end will tend to warm the chair (and the sensor). However it might take a while to cool down afterwards.
    – Nick Gammon
    Oct 29, 2015 at 20:38

2 Answers 2


As has been mentioned in the comments there's many ways of sensing, each with its own pros and cons.

Sensing the presence of something is actually twofold. First you need to know if there is something there, and secondly you need to try and identify what it is that's there and if it fits your criteria.

With passive sensors of the type used by Arduinos it all comes down to a matter of probability.

If you have force sensors, you know when something is there, but you can't work out what is there, only that it weighs a certain amount.

Using capacitative sensors you can get an idea about what the something is made of - how much does it affect the capacitance, etc.

Using ultrasound and/or infra-red beams etc you can get an idea of the kind of size of the object that is in the chair.

So combining them all together you can work out the likelihood that it is a human if:

  • It weighs enough to be a human
  • It has a suitable dielectric response to be a human
  • It is big enough to be a human

Of course, there's always the chance that it's just a big dog, or a sack of potatoes.

So it's all a case of improving that probability by adding more sensors of different types, until you are almost certain that it is a human. You will never be 100% certain, of course.


I've also been trying to figure out a suitable idea for the same question.

  • Image processing would be so accurate in this but there are some problems with this approach, since there is a lot of code work and in case we want to implement this system in restaurant people might not agree.
  • Pressure sensors would work better for the idea but there are some issues with this approach also. Just as mentioned in the question if there is some kind of a weight in the chair there will be an issue. also there are several chairs for a table in case of a restaurant so if we need to know whether the table is empty or not we will have to check whether all the seats are empty or not. In this case we need to pair the chair to the table using some method. we can also take this approach assuming that the chairs of the table will never change i.e. the same set of chairs remain for the same table.

The best approach for this particular scenario would be the sensors being in the table, or having a set of sensors such as a pressure sensor, thermal sensor, IR and verifying the seat is occupied.

  • Also there will be an issue if a device is connected to the chair since it needs power to work, once it the battery is drained you will have to charge it too Jul 6, 2016 at 6:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.