Indoor Positioning - Opportunities and Challenges
You are in a Walmart store to buy a comforter. You are not sure where comforters are stocked. And if this store is like a typical Walmart then it's huge enough to not be able to search manually. Now how do you reach to your comforter with least hassle? Of course, you could seek directions from an employee of Walmart. Or Walmart could put up signs. However, these options are scalable only to an extent. During peak shopping periods, it is difficult to get hold of employees. And, possibly, a store can't put up signs for every product (or category) that it stocks.
What if you could pull out your smartphone, open the Walmart app, search for comforter aisle and hit "Get me there"! And presto! You get turn-by-turn directions on a map from where you are currently standing (inside the Walmart store) to where comforters are stocked! It doesn't matter if it's peak shopping period or if the store has large number of products, this solution will give you the best experience and will work always (well almost!). The store gets the benefit of reduced cost of assisting customers in these scenarios, and providing better shopping experience (the same mobile app can be extended to cover many other scenarios. E.g. with the user's consent it can monitor what the user is searching for. And can run analytics to create collective intelligence and reposition most searched or least searched products)
While the idea of this kind of navigation is prevalent in the "outdoor" world, there are numerous challenges for similar implementation in indoor. Here are some of the key ones:
- Finding your current position - GPS does not work indoors; location from cell tower is very imprecise (it is typically off by at least 100 meters). So, how do we find your position with acceptable accuracy?
- Plotting the location of comforter on the map - in the "outdoor" world services such as Google Maps do this for us. Achieving this in indoor will require us to create full blown maps (possibly geographic) of the internal layouts of the stores (or any other enterprise)
- Giving turn-by-turn directions - this extends from the previous challenge. Giving this kind of directions will require the system to understand the entire layout, from the aisles to walkways to restrooms and everything else. And this knowledge would be store (or enterprise) specific. A solution built for Walmart can't be deployed as-is for, say, Infosys. For the latter, a new layout needs to be built (probably requiring to understand cubicles, conference rooms and other details)
While there are challenges in implementing this kind of solution, solving this problem with "acceptable levels of accuracy" and converting it to an IP has humongous commercial potential; esp. owing to its implementation possibilities.
As part of Infosys Location Based Services (LBS) platform built in Infosys Labs, we are currently working on solving this problem and enrich our platform with indoor positioning as well. In further blog posts we shall elaborate, in more detail, technical challenges, design considerations, and respective solutions (where applicable).