-1

I'm a software developer for quite some time and now I'm going to start developing for Arduino. My question is related to technology I can use. It's also my first time posting here, so if this site is not the proper place for this questions would you tell me which one I could use to get this answer?

My problem:

I need to identify items within a certain range (~3ft). So I thought about a RFID tag on each item and a reader. The problem with that approach is currently the price. The lowest price I found for RFID tags are about US$0.09 per unit. Which is way too much for this project. up to US$0.01 per unit would be acceptable.

Is there any other way I could identify these items, based on range, not being directed to the reader's laser, that would be cheaper?

I though also about something like stores do with their products to prevent theft. Adding a plastic thingy on their clothes and those totems alert if an item with that plastic thingy passed through.

I appreciate any help :)

Thanks

EDIT

Imagine a bag and products being thrown inside this bag. In my mind I'd have a mechanism in which there would be like a rim around this bag that reads the products information. That rim would be a RFID (just for illustration's sake) reader.

Items would be thrown inside this bag, sometimes plenty of items at once. And they all need to be identified.

Something like this:

enter image description here

I'm not actually going to implement this solution on a bag, but I think it perfecly illustrates my point.

I hope I was able to clarify the problem now :)

eestein
  • 99
  • 3
  • The "plastic thingy" uses RFID. – Ignacio Vazquez-Abrams Dec 03 '16 at 20:42
  • @IgnacioVazquez-Abrams cool :) so I think I'm on the right track here :) Would you know of any alternatives to RFID in this case or at least cheaper RFID options? As I said US$0.09 won't work. Thanks. – eestein Dec 03 '16 at 20:43
  • 1
    How about a barcode reader and print barcodes on labels to stick on your things? – Majenko Dec 03 '16 at 23:30
  • 1
    Print QR codes on the objects and use a webcam to read them. PS 3ft is way to far away for rfid anyways. – Gerben Dec 04 '16 at 09:28
  • @Majenko the problem with that is I'd need to directly scan it, wouldn't I (at least AFAIK)? and as I mentioned that's not something that can be requested. Once the item goes through one specific area it needs to be identified. Thanks – eestein Dec 04 '16 at 17:08
  • @Gerben the problem with that is I'd need to directly scan it, wouldn't I (at least AFAIK)? and as I mentioned that's not something that can be requested. Once the item goes through one specific area it needs to be identified. Thanks – eestein Dec 04 '16 at 17:09
  • What do you mean with "directly scan it"? – Gerben Dec 04 '16 at 19:28
  • @Gerben well, using a scanner directly on the QR code, right? The only "variation" of QR codes I know work like, but if there are others please let me know... As I said I need to work on a range basis (like RFID) not on the user scanning an item. Imagine a conveyor belt with the factory's irregularly shaped products rolling over. The QR code might be facing up, sideways, down, etc... there's not way of knowing, so in that scenario I'd need something to identify the items by range, not "directly scanning" its code. Thank you. – eestein Dec 05 '16 at 10:49
  • It would be helpful if you were less vague on the application you want to use it for. You solve the upside-down problem, by having more than one QR code, or by having multiple scanners, or some way to flip the item, so you can check both sides. You don't have to put the scanner near the QR code. Just use a webcam and some computer-vision software. That way the QR code can be anywhere on the conveyor. – Gerben Dec 05 '16 at 15:51
  • @Gerben I understand, I'm not looking for ways to fix the item's placement on the conveyor belt, I know that could be arranged in other ways. I need something to identify a product without being required to directly scan it. Please check my updated answer, hopefully it will make things clearer. – eestein Dec 05 '16 at 16:48
  • I'd place a few cameras around the perimeter of the bag-entrance. No need to "directly scan" it. – Gerben Dec 05 '16 at 19:04
  • @Gerben Yes, but you still have the case in which two products thrown inside the bag at the same time have the QR code facing each other, something that would prevent the cameras to read it. Thanks. – eestein Dec 05 '16 at 22:39
  • How about using NFC? – duck Dec 06 '16 at 08:15
  • @dhimaspw Thanks, I though about that, but the prices I found are even higher than those of RFIDs... I tried most Chinese sites and the prices are higher, unfortunately... – eestein Dec 06 '16 at 13:29
  • Is there a way to re-use the RFID tags? Or maybe transport the bags in a cardboard boxes (with a label), so the box itself can be scanned? – Paul Dec 06 '16 at 13:44

1 Answers1

1

Seems like you're stuck in "requirements hell". In my theorie this means that your requirements are too strict, and you or your client won't let go of these requirements.

It can help to do a research, set up some "research questions". And if possible with some sub questions, like:

"How can we measure multiple objects at a price of only 0.01$ per object. Withouth having direct line of sight to all sides of the packaging."

With sub questions:

  1. What techniques are available to identify objects.

  2. In what environment will this system work?

  3. How do other companies detect Products?

  4. What techniques fit these requirements and environment.

And if none fit at question 4

  1. How can we change the requirements or environment to have one technique fit.

The idea is, that after the research, you probably have a valid answer. But also that you can convince your Boss/client (or yourself), bases on your research, what solution is acceptable, even if it means altering the requirements.

You may also want to check how other companies do this, since it's a pretty common thing to identify objects etc.

For us, it is extremely difficult to say whatever will fit your needs. And you'll most likely have to adapt something.

To start off, you will either need some kind of vision or RF. But any RF tag is most likely to be too costly.

So you're either stuck with vision or increasing the price limit.

With vision, usually you either require some kind of barcode or qr code. Or you'll have to detect the object itself.

Another wat could be to differentiate on the objects' variables. Like check the weight, color, conductivity or even shape. This could be tricky, but not unlikely.

Mind that it also depends on what items you want to detect.

Paul
  • 2,703
  • 2
  • 21
  • 40
  • Well, I agree that my question was lacking information, but research was done and that's why the RFID solution was commented here. After much research RFID was found to be a suitable solution. The problem is the price. The requirements are there, I didn't specify them all when I first posted, but they are there now. So even though RFID does solve the problem, it is at a too high price per item. That's why I'm looking for solutions with the same capability that are cheaper. Thanks for your answer. – eestein Dec 06 '16 at 13:26
  • Well, if there is a price restriction, how comes that RFID is found as a suitable solution? Since, it's not a viable solution, based on price. – Paul Dec 06 '16 at 13:34
  • But, my point is, that you should check out the options you have and present this options to whoever wants you to develop this. So the conclusion of your research will be: "You either have to pay a few cents extra and get RFID working OR you can use vision (using QR or something specific), since it's cheap but will require the label to be visible. – Paul Dec 06 '16 at 13:37
  • Paul, thank you for your comments. RFID was found to be a suitable solution technically, when it came to price it was rejected. The only reason why I said that was to explain what was needed for this project (RFID suits the needs). As to "paying a few cents extra" is impossible since it would render the project inviable. Answering your other comment, that is not possible as well. I'm going to try to use photoshop to draw exactly what I mean, I think I'm not being able to explain myself properly. Do you know of any tools to prototype I could use? I'm gonna start drawing on PS for now. Thanks. – eestein Dec 06 '16 at 14:28
  • You could describe the environment a little better. But, if RFID is the only technical possible solution, but it's not economical viable solution, it means that there is no solution ;D – Paul Dec 06 '16 at 15:36
  • My point being that your requirements have to change (or find a more practical approach, but that would require an awful lot of knowledge into the environment and probably even reshaping some requirements) – Paul Dec 06 '16 at 15:38
  • Well, yes, but I never said it was the only possible solution. Just that it was one that we found... hence why I'm here looking for a cheaper technology with a functionality similar to RFID. I'm going to update my post and explain in details what I need. :) I think now the example will be clear... :) – eestein Dec 06 '16 at 19:08
  • If there was something like RFID, but 10x cheaper, it would've been more common than RFID :p (or actually, nobody would use RFID, right?) I have no other solution as altering the requirements (by proving that the current requirements are too strict). And if that's impossible, you should reconsider your project. Trying to do the impossible sound honorable, but in practice it'll get you nowhere. – Paul Dec 07 '16 at 00:59
  • And I get the idea of detecting items that are being put in a bag. But you don't tell why you want to identify these? Are you making an "automatic shopping cart" or another way to sort items? I don't see in what context it will be used. – Paul Dec 07 '16 at 01:02
  • I don't have enough reputation here to start a chat, could you? I think it will be better. Thanks anyway :) – eestein Dec 08 '16 at 11:48