I was working with Target when they did this roll-out. As sinister as everyone here seems to make it sound, I can assure you at that time it was not. If you're familiar with the inception of corporate projects, this project was no exception to any of the stereotypical fumbles and foibles of any effort of this scale.
It began simply because Apple said "...iBEACON.." and everyone corporate wanted the new buzzword in their portfolio to let people know how hot/hip/tech they were (toys -r- us considered it for awhile) so they could pull the kids away from the internet and back into brick-and-mortar. How can we use these? What are they good for? No one really cared, they just wanted them in the store and we were there to sell them that (at a premium).
In 2015 - indoor location was bogus. Everyone in this forum seems confident that there are multiple ways you can determine location with Wifi (round trip packet time) or bluetooth (RSSI). A cacophony of radio in a catastrophically noisy environment does not work to provide reliable location information.
In fact - it was so bad, that there were a handful of other equally unrealistic solutions being pedaled by everyone from universities to light-bulb manufacturers. One such solution was to profile the accessible space of a building using a phone's compass, and then use the observations from a client device compass to identify how generated patterns correlated to the profile for the building.
The torrent of data you see pouring from your phone to the service via wireshark is real-time sensor data that is being used to feed a service side bayesian / markov-chain / monte carlo / kalman-particle filter / keyword soup monstrosity trying to generate some possible marginal confidence in a probable location. We found that this system was most effective in turning your phone into a pocket warmer, but sold like hot-cakes in a B2B setting wherein the intended end user had absolutely no voice.
Corporate wanted to buy it so they could sell advert-space (pop-up coupons) to affiliates. So we sold them something that worked barely well enough to provide a one-popup demo to potential interested parties. The affiliates bought the magic, corporate paid our company an arbitrarily large quantity of dollars for the service.
A deployment of thousands of coin-cell driven beacons per store, placed within reach of bored youths, maintained by an underpaid associate staff is of only questionable utility.
On a scale of things to worry about, ranging from nuclear holocaust to e. coli in your produce - this ranks a solid -3. In fact, in the 4 years since I've worked in this field I think the only advancement that has been made is that it's harder for third parties to sell because no one can do it well, so why not just do it in-house? It's cheaper and has the same garbage result.
If you think I'm wrong - go do it yourself. All these signals are easily grep-ed within any store (there's no way to make it proprietary), and you can create your own model and out-sell the proprietor. Surprise me. With an actual, viable client-oriented product (and assuming users actually want reliable indoor location) you'll make bundles selling it to the valley. Everyone is trying to do it (even cisco tried for awhile) - no one has.
Michael Kwet has read all the marketing copy, and rewritten it for popular consumption as a product. The insight/value provided by these systems is far more sparse than implied.
Bluetooth 5.1 promises a sub-meter accuracy in distance finding using Bluetooth beacons. When those devices become available you'll probably want to re-evaluate those statements.
It began simply because Apple said "...iBEACON.." and everyone corporate wanted the new buzzword in their portfolio to let people know how hot/hip/tech they were (toys -r- us considered it for awhile) so they could pull the kids away from the internet and back into brick-and-mortar. How can we use these? What are they good for? No one really cared, they just wanted them in the store and we were there to sell them that (at a premium).
In 2015 - indoor location was bogus. Everyone in this forum seems confident that there are multiple ways you can determine location with Wifi (round trip packet time) or bluetooth (RSSI). A cacophony of radio in a catastrophically noisy environment does not work to provide reliable location information.
In fact - it was so bad, that there were a handful of other equally unrealistic solutions being pedaled by everyone from universities to light-bulb manufacturers. One such solution was to profile the accessible space of a building using a phone's compass, and then use the observations from a client device compass to identify how generated patterns correlated to the profile for the building.
The torrent of data you see pouring from your phone to the service via wireshark is real-time sensor data that is being used to feed a service side bayesian / markov-chain / monte carlo / kalman-particle filter / keyword soup monstrosity trying to generate some possible marginal confidence in a probable location. We found that this system was most effective in turning your phone into a pocket warmer, but sold like hot-cakes in a B2B setting wherein the intended end user had absolutely no voice.
Corporate wanted to buy it so they could sell advert-space (pop-up coupons) to affiliates. So we sold them something that worked barely well enough to provide a one-popup demo to potential interested parties. The affiliates bought the magic, corporate paid our company an arbitrarily large quantity of dollars for the service.
A deployment of thousands of coin-cell driven beacons per store, placed within reach of bored youths, maintained by an underpaid associate staff is of only questionable utility.
On a scale of things to worry about, ranging from nuclear holocaust to e. coli in your produce - this ranks a solid -3. In fact, in the 4 years since I've worked in this field I think the only advancement that has been made is that it's harder for third parties to sell because no one can do it well, so why not just do it in-house? It's cheaper and has the same garbage result.
If you think I'm wrong - go do it yourself. All these signals are easily grep-ed within any store (there's no way to make it proprietary), and you can create your own model and out-sell the proprietor. Surprise me. With an actual, viable client-oriented product (and assuming users actually want reliable indoor location) you'll make bundles selling it to the valley. Everyone is trying to do it (even cisco tried for awhile) - no one has.
Michael Kwet has read all the marketing copy, and rewritten it for popular consumption as a product. The insight/value provided by these systems is far more sparse than implied.