Proximity triangulation with (BT/RF/WiFi/?)...

New to MCUs & Arduino, lots of electronics experience, though.

I want to make an app for, say, Android that can triangulate a user's position based on the signal strength of beacons, or nodes. The app would then compare the relative strength, and id, of the received nodes, to a known table, and thereby know where it was in the store. Outside cell coverage cannot be used, so everything must be done within the retail store.

I was hoping that the nodes/beacons in the store could double as the network the phone would connect through to transmit it's location, customer paths, then receive info about that location (as an example), but I ran into a snag.

When I got my UNO to drive an HC-06 and connect to it, I realized that BT may not work. The each pair of BT devices creates an exclusive pairing. This is far from a mesh, and makes triangulation impossible. If I could just sense the signal strength of each-without pairing, that would be fine, but it leaves me with no communication vehicle.

I've got some RF modules coming to test with as well, but I'm really trying to avoid TCP/IP over 802.11 due to power requirements. I was hoping to power these with solar (store lighting) & rechargeable batteries, not running extension cords everywhere. Like a magnet on the bottom, panel facing up?

Use Case

  1. Imagine there are 3 of these critters for each aisle (endcaps & middle). Where you put them really isn't precise, as long as you can receive 3 signals at once, you should be able to triangulate.
  2. You take an admin-version of the app, and go about the store. When you are standing in front of something notable, you mark the position in the app (it is using signal strength & node IDs), and make note of whatever it is you are standing in front of.
  3. If for customer pathing, it could be as simple as end of all aisles and the middle of all aisles. This, taken with times, would tell you where customers are wandering around in your store.
  4. The nodes facilitate communication between the cell phones in the store and, say, a desktop PC in the main office.

If I could find a way to power the nodes AND allow them to mesh together (connect to more than one at once), I could knock out 2 birds: The beacons would double as the network nodes. My reading tells me that keeping a busy WiFi card running is a large power drain, but I'm not even sure if WiFi is needed...

Since this is Project Guidance, what are your thoughts?

Thanks!

I'm not sure that what you have described is at all feasible. Even outdoors where there are no obstructions signal strength would be a very imprecise indication of distance. Inside a store I suspect the distances would be too short to make a reliable difference in signal strength even if there were not metal shelves and packaging to cause reflections and to absorb the signal.

I think you would get much better results with some sort of overhead camera system - but that is far beyond the capabilities of an Arduino.

...R