I am trying to design an employee tracking system, to track employees from room to room, for saftey purposes.

Somewhat obviously, if requires each tracked person to carry a device which uniquely identifies them.

Which will obviously have to be battery powered, with enough battery life to last a full working shift, let's say 12 hours?

It is becoming apparent that I am looking at Bluetooth Low Energy – unless someone can make a compelling argument for WiFi, ZigBee., etc.

Please, bear in mind that the device and battery must be worn in an industrial environment, so should be robust, and also easy to charge when needed – at worst overnight, ready for the next shift.

The device need only identify itself, as it passes various access points (presumably by pairing with them, or offering to do so?)

Looking at the List of Bluetooth profiles , I am guessing at the Proximity Profile (PXPP) ?

In any case, what I do not want – close as it is - is a cheap sports/fitness band. What I would like is something very like that :

  • Rubber wrist band – nice to have, but not mandatory
  • Rugged & robust – is quite important
  • Button/cell battery – for small form fit
  • NO built in apps, e.g. health / fitness
  • It would be nice if I were able to load my own software. For instance, to vibrate the band remotely, or display messages. What I don’t want is lots of built-in fitness tracker stuff, which is not associated with my system

The Flora Wearable Bluefruit LE Module looks good, but

  • it’s not totally off the shelf : “Sew 4 traces (or solder 4 wires)”
  • There is no indication of battery life (although I don’t expect any problems with BTLE)
  • I would prefer something worn on the wrist, but could live with this

On the plus side it seems easy to program “After you connect to the Bluefruit, you can send commands wirelessly in under 10 minutes”. Also, “Built in over-the-air bootloading capability so we can keep you updated with the hottest new firmware. Use any Android or iOS device to get updates and install them” seems a big plus.

Is there any other cheap (preferably sub $50) wrist wearable which is programmable, preferably with a snap-in coin/cell battery?

[Update] I now have a strong preference for BT 5, rather than v4, but can't seem to find any. The reason being that the BT 5 spec has provisions for proximity, which could save me a lot of effort on triangulation.

  • 1
    See this company: litumiot.com/employee-people-tracking Commented Feb 23, 2019 at 20:35
  • Looks great, and is pretty much what I was planing to develop. It uses RFID, though, not BT. I could live with that, although it obviously has to be active RFID, not passive, which dries up the price of the readers. Hmm, I just read further and their readers will also detect BT-LE. I will contact them, as their underlying infrastructure interests me, but I really want to develop the software which acts upon employee location myself (plus maybe also transmit to BT). A great comment - feel free to post it as an answer - but I am still looking for the wearable :-(
    – Mawg
    Commented Feb 24, 2019 at 8:05

3 Answers 3


I feel that Hexiwear is a pretty good platform, although a little expensive. I feel it ticks most of your boxes:

  • rubber wrist band - check
  • rugged and robust - not really, but it's enclosed in that wristband
  • small form factor - check
  • no builtin aps - check
  • loading own software - that MCU is fully programmable
  • sub 50$ (49$) to be exact

You might initially be thrown back by the small battery (190 mAh) but:

  • that's all you can reasonably get in such a small form factor
  • it's enough to get you through the day if you program it well

Regarding deep sleep - in such an application it is pretty much required, not an option to be explored.

As a bonus you have a pulse oximeter and heart-rate monitor which could detect some health conditions in your employees (like fainting, there is even a tutorial on that).

  • Thanks, I had previously looked at Hexiwear. I thought it a little expensive, even though I now say sub $50 ;-) I suppose that the screen is a bonus. Otoh, take a look at the RSL10-COIN-GEVB in my answer above. Do you know which languages I can code the Hexiwear in? Ada would be great (it's ARM, so that should be possible). Failing that C. Failing that, Python. A major bonus is that I can debug it properly, so I think that I will buy one and experiment. Thanks for reminding me of this.
    – Mawg
    Commented Sep 27, 2019 at 8:59
  • 1
    @Mawg definitely C, Rust likely as well, Ada should be possible but there would be no libraries. And it has vibrations (forgot to mention in the answer).
    – jaskij
    Commented Sep 27, 2019 at 9:01

I am answering my own question, in case it is of any interest to others.

For development, I have found an excellent and extremely cheap solution. In fact, it could probably be used in non-profits. Only for a commercial product would I need to have something custom manufactured, as I can't find a COTS solution (needs to be smaller, more rugged, with a better battery).

So, for US $14.90, plus postage, I found this:

M5Stack Multi-function Watch! With 700mAh Battery for Arduino & Micropython ESP32 Core Intelligent Programmable Watch with Band

It's slightly chunky, and I don't like only having a 700 mAH battery (but the ESP32 has great power management), but it's more than good enough for development.

enter image description here

enter image description here

Hmmm, the Google cached version of the page shows it as costing less ($10.70), with a batter (850 mAH) battery. Strange days indeed!

Also,this one looks smaller, with a better form factor. I will have to enquire where to but it & update this answer. enter image description here

[Update] from forum.m5stack.com/topic/548/m5stick-watch/9

the original stick can be purchased via the following links: https://www.aliexpress.com/store/product/M5Stack-Official-New-M5Stick-Mini-Development-Kit-ESP32-1-3-OLED-80mAh-Battery-Inside-Buzzer-IR/3226069_32994893797.html?spm=2114.12010615.8148356.13.25ba781bqWZV2X


The main differences between the stick and stick c are that the stick c has a smaller screen but it is color opposed to the monochrome oled of the original, the power management chip is different, there are a bunch more sensors added etc.. for a detailed comparison you can check the docs pages for both the stick and the stick c

https://docs.m5stack.com/#/en/core/m5stick https://docs.m5stack.com/#/en/core/m5stickc

Also costs sub $15

[Update] link to project exploring the deep sleep possibilities of the M5stack watch

  • Hi @Mawg - were you able to get one of these products to work for employee location tracking purposes?
    – siliconpi
    Commented Sep 26, 2019 at 5:26
  • Yes, as prototype, but I am somewhat concerned about battery life. See last line of this answer, where I added an update, also my new answer which looks to be extremely interesting
    – Mawg
    Commented Sep 26, 2019 at 14:08
  • I wouldn't trust anything you can only buy on AliExpress in any kind of professional environment. No guarantees, often no datasheet.
    – jaskij
    Commented Sep 26, 2019 at 21:35

I am still experimenting, but will try to keep this thread up to date.

The M5stack works well enough, and has the advantage of a display - and a microphone (not sure if it acts as a speaker) and does not have built-in vibrate to get attention functionality, although that could be added.

I still need to need to run test on battery life link to project exploring the deep sleep possibilities of the M5stack watch).

Right now, I just discovered the RSL10-COIN-GEVB, which uses a standard wtach/coin cell battery, and am very excited as this page says

The board incorporates a temperature sensor, NCT375, and RSL10, industry's lowest power Bluetooth 5 SoC. With the included CR2032 coin cell and an advertising packet being sent every 2 seconds, the device lasts over 6 years!


The beacon follows Eddystone, an open beacon format from Google. A web URL and ambient temperature data embedded in the advertising packet can be read using Beacon Scanner or other freely available beacon scanning apps from the Android and iOS app stores.

That seems ideal. I am now questioning whether I really need a display. I suppose I could ad done, but what might it do to the battery life? Maybe only power it on for 60 seconds when there is a message? Or use eInk?

I will update this when I have experimented some more. Btw, there is a battery-fee , always on, solar powered version It costs 4 times as much, but sounds well worth it for certain applications

  • 225 mAh give or take... lasting 6 years give you roughly a 2 microamp average current consumption. That's low, very low, and require great attention to how you write your software, often getting into the nitty-gritty stuff.
    – jaskij
    Commented Sep 26, 2019 at 21:33
  • And that device does not have a screen.
    – jaskij
    Commented Sep 26, 2019 at 21:34
  • Sorry for the spam: treat the stuff in the datasheet as the perfect scenario. Anyone working in electrical engineering will tell you that getting the numbers from the datasheet in your device is a achievement.
    – jaskij
    Commented Sep 26, 2019 at 21:36

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.