I did not understand at first your idea was number one. It would be the access points who would locate the user’s laptops and based on their location advertize them self and allow them to connect. This is solvable as proprietary location layer can be added to router by a manufactures that would allow them to do a cooperative triangulation of the users. But privacy legislation will probably prevent you from doing this in Europe and some other parts of the world. I America I even believe some would make interactive advertisement in the shop alleys that would adapt to your credit score as you walk by.

The second side is the one interesting most the users. Where is my product in a shopping mall, where are my friends, what can I visit?… The best examples I have seen in this category with WIFI, is a hotspot in a shop and their homepage pops up when the user connects to it. I know a bookshop where they did this, they recycled their online book search app so you can use your UMPC or cell phone to search books in their database over their WIFI while sitting in the alleys and not having to queue for the computer kiosk.
WIFI could help in non precise indoor location: just let the admin be able to include 3D GPS coordinates of the location of the router in its beacon. A reasonable estimate of your location is them the coordinates of the closest hotspot.

For robots there are much more issues, not only they need to know where they are but they need also to find out them self what is around them (collision avoidance). But this is not the urgent problem, the real problem is on how to make them do the dishes and the laundry.

A spread spectrum approach is probably the only way to really assure a reliable system in the varied and reflective world we live in. Depending on the desired resolution and the practical indoor range required you’re talking about certain frequencies/signals too. So implicit to my post, WIFI as we know it would change too. That’s not a bad thing. Anyone that’s tried to assemble a WIFI network with lots of access points can attest to this. As I mention, the WIFI-based mesh would then also double as a location based system for other devices. That could be very interesting. But I’d still go back to the motivation that it could help improve the public WIFI issue by simplifying the experience for the user.

In terms of your super market example, yep, that’s where something like this could really sing.

From a Robotics angle I’ve worked on a variety of indoor positioning systems, some radio based, some IR based, some laser based, some visual based. It’s a hard problem. However, when someone figures it out, I think it’s going to be a game changer.

Right now, I’m thinking the “solution” is a radio-based one, made practical in silicon, and integrated with WIFI to enable broad value.

For location determination I believe more in GPS systems with an inertial system. Wireless signals especially Wi-Fi signals are just too unreliable. Or for a robot why bother with Wi-Fi, just use it as any other fixed signal source.

For Restricted geographic wireless zones, you can:
-Use the 60 GHz range spectrum. It’s absorbed by the air so it radiates only a couple of meters.
-Use a frequency that is absorbed by building construction material: wood-metal in US, concrete-bricks in Europe and Asia.