We have been recently making some references to the WiFi-based indoor positioning system currently under development by the SCYP team. Over the past few weeks, we have been wrestling with the system's design and tackling some of the initial barriers to the implementation of our very first prototype. But thanks to Renaud, a committed open source advocate, we have just solved a simple but necessary step for the success of this project.
See, a few years ago, Renaud wrote a program that converts CAD files in DWG or DXF format into more standard SVG images. We used Renaud's program and other open source tools to convert all campus blueprints into standard image files so that they are easily accessible. These image files can in turn be used to build a database of associations between campus locations and WiFi scans (i.e. fingerprints). The principle is simple: with blueprint in hand, we can go to every building on campus collecting WiFi fingerprints as we walk around. For each fingerprint collected, there will be a corresponding location in the blueprint. Then, when sufficient associations have been made, and with the help of a very fast and smart algorithm, we should be able to tell the position of a mobile device just by looking at the WiFi fingerprint, and thus, enable indoor positioning just as with GPS!
Now, besides the fact that Renaud's program worked like a charm even a few years after he wrote it, he was also happy to make some minor changes that allowed us to easily extract the dimensions of the original blueprints, which is necessary to define real coordinates from our converted images. So, hats off to you Renaud!
No comments:
Post a Comment