0

I apologize if this question is inappropriate for this forum, but I'm coming up short in my research.

The problem is as follows: I would like to find a network of devices capable of tracking people moving between key areas of a building, and capable of integrating with a REST endpoint upon proximity triggers. E.g. when any person passes a wall-mounted reader within 5-10 feet, an http POST can be sent to a local server.

I had envisioned being able to accomplish this task with RFID-tagged ID cards and large-proximity wall-mounted RFID readers with some kind of USB-ready software. However I was unable to find such a system with my google-fu.

Does anyone have any experience in this general area, or have an interesting thought on a potential home-brew solution? I welcome any and all advice, thank you very much for your time!

Will Buck
  • 101
  • 2
  • 2
    Related: http://electronics.stackexchange.com/questions/30895/ – stevenvh Jun 12 '12 at 17:41
  • Since most people have phones and most phones have Bluetooth, I would go that way. Unless you can put the RFID on something they already carry, such as a corporate ID badge. – Brian Carlton Jun 12 '12 at 17:42
  • Its actually a pretty old-school crowd with low smartphone penetration, but they are required on site to have a badge, which I was hoping to find an RFID solution for (replace the badges with some that have RFID embedded, or adhere / cover the existing badges w/ a passive signal emitter) – Will Buck Jun 12 '12 at 18:09
  • 1
    @BrianCarlton one problem with Bluetooth is I think you'll need to get the user to agree to pairing. A bit of a barrier. RFID ID cards are probably the best. – kenny Jun 12 '12 at 18:44
  • also related: http://electronics.stackexchange.com/questions/28065/indoor-location-positioning – clabacchio Jun 13 '12 at 07:21
  • Face recognition software. The software has evolved so well, that now it may the most practical solution. There is even a hardware level support for this, say in some camera ASICs, FPGAs or firmware. –  Jun 13 '12 at 22:50

5 Answers5

2

Systems such as this one use actual wrist worn transmitters with proximity constrained receivers located at various points in the building, typically at doors and elevators. The receivers are configured to detect a wrist worn tag when it is within about 3 meters. In your case a system like this could get expensive quite quickly (for a given value of 'expensive'...) since you'd want receivers in many places. It might give you some ideas though. There are several companies that make similar systems and target the same market. None of these have quite what I would describe as a network back-end as you describe, but that's the easy part. (Speaking as a software developer who has recently integrated one of these into something very much like this.)

The RFID badges that I'm familiar with need to be within about 10cm of a sensor to trigger it, and are expected to be used intentionally, i.e., to unlock a mag-lock and allow access to a room, etc. My impression is that you'd like to track personnel movement without requiring the personnel to have to think about it.

MickeyfAgain_BeforeExitOfSO
  • 542
  • 2
  • 9
  • You are exactly right in your second paragraph, and that was what I was afraid of hearing. The people to track are actually customers, there are perhaps a thousand of them at any given time, and the only thing they are issued by the building is ID badges right now. I will have to look into the cost of those wristbands and see whether that may be an option. – Will Buck Jun 13 '12 at 22:22
2

What you are describing sounds a lot like what Radianse makes. Check out the general system diagram. The system basically consists of small battery-powered RFID tags, receivers at known fixed locations throughout the building, and software on a server that puts it all together and figures out what's where. This is dumped into a database that applications query and can get real time notification from.

By the way, I designed the receiver shown in the middle of that picture and wrote firmware for some of the tags.

Olin Lathrop
  • 310,974
  • 36
  • 428
  • 915
  • This is definitely closer to what I'm looking for, I will have to take a look into the cost to see if it makes sense for my problem space. Thanks! – Will Buck Jun 13 '12 at 22:27
1

I am not aware of any commercial solutions, but this is a typical application for sensor networks.

One of the nice examples is TI Chronos [ http://processors.wiki.ti.com/index.php/EZ430-Chronos ], which is a kit consisting of wrist watch with programmable radio, and a USB transceiver to go with it. In the simplest case, you'll program watch to transmit a beacon every few seconds (perhaps with with reduced power -- default settings will give you ~100 feet indoors), and setup receivers everywhere. You'll want to have small Linux boxes with ethernet port and USB port for each receiver to receive beacons and translate them to POST requests. Alternatively, use peoples' existing computers (this worked great at university/lab, where we could just ask people to plug the receiver into their work PC).

If someone had lots of time to spare, she could use real sensor OS such as Contiki or TinyOS and setup a proper mesh. This will significantly cut down on number of computers used -- the receivers would forward messages to a single "sink" node. That would save on cost of small Linux boxes.

Alternatively, if one had hardware development experience, she could design receivers to hang from single RS485 bus. Or make transmitters smaller -- there are lots of options with sensor networks.

theamk
  • 3,128
  • 11
  • 15
  • thanks for the thoughts, I don't think this would fit the scale of people I'm needing it for but I appreciate the effort! – Will Buck Jun 13 '12 at 22:28
1

Maybe it's not just the solution itself but I find it quite hard to solve this with RFID. I mean solve it in a cheap way. The RFID readers that can read the tags from meters are really expensive and I think an authentic solution would let the people just walk where they want and not require them to show their tags every step.

Szundi
  • 215
  • 1
  • 6
  • Exactly. The goal is a cheap thing people carry around with them without even really realizing (ID card etc) and a hallway-sized range for a reader to be setup in key bottlenecks of a hallway system to know what 'areas' people were last seen in. Glad to hear I'm not the only one having difficulty! – Will Buck Jun 13 '12 at 22:30
0

If the building is relatively large, you could set up WiFi hotspots in different sections of the room, and assuming everyone carrys a smartphone, detecting the SSID of the router its connected to would give you a general idea of where everyone is.

Jonny Lin
  • 101
  • 1
  • Good thought but as I elaborated on in the comments, really isn't a viable option for me, this crowd probably isn't 50% penetration for smartphones. Perhaps this would work for someone else stumbling upon this question though, thanks for the input! – Will Buck Jun 13 '12 at 22:37