6

Im thinking of developing augmented reality app for my school library. Basically we alreayd have a machine which will give us the duis decimal code for where the book is but the librarian still has to tell us where this is. An augmented reality application on top of this to who people where the book is might work well I think.

Just wondering how difficult it is to do such a thing without GPS as its indoors.

Intel had a good solution http://www.youtube.com/watch?v=WciFx66ojA4 but is it a viable app to make?

OVERTONE
  • 11,797
  • 20
  • 71
  • 87

2 Answers2

4

The IMHO easiest way would be to place some WLAN senders in the library and the use them to calculate the position based the signal strength of all WLAN senders in reception range.

There are a couple of scientific papers and demo-installations available around the world:

http://scholar.google.com/scholar?q=indoor+localization+wlan

Robert
  • 39,162
  • 17
  • 99
  • 152
  • Is WLAN triangulation really reliable? If he wants to lead people to a specific book the algorithm has to be very exact, so the user won't turn right one meter too early. I could imagine a combined algorithm, where the approximate position is determined via wlan, to guide the user near the shelf and then use marker-based tracking to determine the exact position. This would be quite an effort though, if you're going to store relative positions of all books to a certain marker. – Pedro Oct 19 '11 at 18:05
  • WiFi positioning coupled with Sensor data for [detecting Inertia](http://www.starlino.com/imu_guide.html) should improve the accuracy to some extent. – Reno Nov 16 '11 at 04:46
3

If you haven't worked much with Android before, specifically with capturing and overlaying its video stream data I would say that creating an augmented reality application would be very difficult. I worked on a project capturing and analyzing live video streams through android for on-device analysis and found it significantly challenging.

If you're looking for an alternative I would suggest a static mapping application that showed a blueprint-like layout of the library, your current location at the terminal that provides the Dewey Decimal listing, and the calculated location of the book they just looked up. You can even add a section about the Dewey Decimal classification system and become your librarian's best friend.

If you really want to go for live navigation indoors your best bet may be through a combination of the compass and accelerometer data so you can infer direction and distance as a user moves through the environment. There are some good papers around in regard to this but you may need to go through the IEEE or ACM libraries to look them up.

Carth
  • 2,303
  • 1
  • 17
  • 26
  • 3
    Totally agree with Carth. Indoor navigation is a complete pain right now. Most examples use triangulation based on signal strength, but that method is error prone when you're around objects that block the signal. Dealing with multiple floors is another problem entirely. The 2D map would work well, but obviously not as cool as AR. Using Compass and Accelerometer does work, but only over short distances becaues the accuracy is low and drift occurs. You can account for drift by having reference points that "reset" the device to a known location. QR tags could be used for this. – Jared Bienz - MSFT Nov 02 '11 at 18:19