* Minimize BOM cost. Only 1 in N lights would actually have the full sensor complement, so they needed to communicate.

* Minimize installation cost. They just wanted to plug into a light socket, not run network cabling.

* Push data logs to a central server. They didn't want to send a tech physically to each lightbulb to get data for e.g. energy usage certifications.

plus other obvious requirements.

All of that made it really easy to just stick a beacon tag inside employee badges and measure the RSSI from the mesh lightbulbs (since they already tracked that to discover who their physical neighbors were). Instant employee monitoring.

Are these hallways for staff only?

Not as far as I was ever aware. They were intended for guest areas.

So why is the RSSI thing worth having?

RSSI is just signal strength. Tracking it allows you to approximately triangulate where people are so you can say "ID #xxxx was closest to the first floor east stairwell at 10:02am". Maybe nothing is done with that data, but it's ultimately a surveillance system and enables all the potential abuses that can entail.

I know what RSSI is, it seems like complete nonsense even pretending this is for energy efficiency if you have to keep track of where the non-staff are anyway - or is the people-tracking just an optional extra?

I think people tracking was just an optional extra/add-on. It sounds like the RSSI signal was there in the first place to set up the mesh based on what other lights had active sensors...

Because by measuring multiple RSSI you can trilaterate and track employees relatively accurately. If an employee lingers somewhere unusual, you can have a manager ask them why.

That isn't what the system was notionally designed to do

Yes. The commenter is upset because a mesh network that he helped build for another purpose also provides a convenient dense network of radios that can be used to track staff, and that this is now the primary use case.

"All of that made it really easy to just stick a beacon tag inside employee badges and measure the RSSI from the mesh lightbulbs (since they already tracked that to discover who their physical neighbors were)."

I assumed the rssi thing was there from the start, apparently mistakenly.

If it's any consolation to the original engineer it feels like a non-trivial thing to add

Most nodes in a radio based system will do rssi measurements to any other nodes that they need to communicate with directly, as part of deciding appropriate tx/rx amplification levels / when nodes are unavailable / etc. These functionalities are often easy to access to enable easier debugging, so it's a relatively straightforward change to start using them to scan other things on the same protocol etc.

The original engineer states the RSSI thing from the start because it was an input to the meshing algorithm.

Using it for location of employees was the new part.

By the time they got to "just stick a beacon tag inside employee badges and measure the RSSI from the mesh lightbulbs" it's pretty clear that it's one of the features they had in mind, though. You wouldn't "just" re-issue all your employee badges and spend engineering time on integrating it like this if tracking the employees wasn't the goal.

For some additional context, the original feature genuinely was energy efficiency. <Hotel chain> was trying to get some green certifications for a flagship and went looking for partners to do the actual efficiency stuff. That company found a sub, and so on until I was contracted to do the actual firmware. Everything got done and it was installed in a test hotel in Vegas.

After that initial success, one of the intermediary contractors came up with the surveillance idea (among others) to try and find reasons for <hotel chain> to roll the system out to more facilities.

Guess you used wireless communication to avoid cables then also to push logs and communicate with other lights.

I am starting to think the requirements were crafted to push towards enabling surveillance without outright stating it.

PIRs are cheap and last basically the lifetime of the hotel. I would need to see some actual data to believe this whole project didn't actually cost more than installing a bunch of PIR lamps.