I am currently conducting research that attempts to implement Gaussian Processes for Signal Strength-Based Location Estimation in a distributed grid environment using parallelized algorithms of GPs that will be consumed on an Android application. While I do not have a working implementation yet, I will certainly keep those interested updated. And you are of course encouraged to venture down this rabbit hole yourself.
The one thing I have discovered during my research is there is no real platform, commercial or open source, for indoor navigation. What I define as a platform is a collection of appropriate algorithms for determining accurate indoor location, floor plan base maps (and additional data specific ones), a cloud or grid environment for location computation (you don't want to be doing those computations on the mobile device, and you should plan on scalability for the future if you do actually figure out how to do this - it will be in high demand), and either a mobile application or API that supports the above infrastructure.
Your best bet currently for an indoor navigation platform is to leverage existing players in the market - Google Maps, Bing Maps, ArcGIS - and overlay your own base maps for the indoor locations of interest. The accuracy of the location indoors depends on which location provider(s) you use (A-GPS, cellular networks, WiFi). What mobile OS are you wanting to use? If you are using Android, they have an excellent primer on obtaining user location, and I have a public GitHub repository that allows you to log location observations from each location provider for later analysis. DISCLAIMER: That application was written as a supplemental component to a research paper, and it is by no means complete or fit for anything. ;)