Indoor Location Positioning

Question

We are developing a mobile application for smartphones (Iphone, Android, BlackBerry, etc) and want to allow navigation at indoors.

We can buy and place indoor gps antennas but will this solve our problem and let us to navigate users?

I searched a lot and found some systems; but they are bundled with chips and antennas. But we couldn't attach chips to users phones.

So do you have any suggestions?

score 5 · Answer 1 · answered Mar 14 '12 at 21:07

Indoor GPS antennas won't help because they are just repeaters for the GPS antenna on the roof. There will be propagation differences between the roof antenna and the repeaters, but these won't help resolve actual location differences. The purpose of such antennas is to allow devices in the building to locate to that building and to provide the GPS time signal.

Indoor location is tricky business. No single solution has emerged, although there are a number of companies trying to serve that space. Some use WiFi signal strength, signal strength from active RF tags, carrier phase from active RF tags, ultrasound, IR, combinations of these, and probably more I am not aware of.

thanks for information especially indoor gps antennas; because i thought that could help us :( yes as you sad, lots of companies trying to find a solution but most of offer to sell a chip to track device at indoors. but we have restrictions about this, this place is a public area and we couldn't plug-in that chip to their phones. — bahadir arslan, Mar 15 '12 at 06:54

score 3 · Answer 2 · answered Mar 14 '12 at 22:17

3

Could you use special filters on the data from accelerometers included in most smartphones? I would imagine that people walking around indoors have very specific patterns (i.e. it's unlikely they're moving faster than 3mph, except in very short bursts. Its also unlikely they're moving at any speed between 0 [exclusive] and 0.1mph for very long). If you take enough of that data and push it into a neural network, you may be able to get a fast way to filter out electrical noise and drift. It might drain the battery pretty quick, though. Maybe you just have to wait until there's a standardized protocol for indoor navigation? Or you could make one... no wait! Dibs! I'm gonna make that. Patent. Copyright. Sued by Apple. Then by Google.

answered Mar 14 '12 at 22:17

Mutant Platypus

49
1

+1 lol ... Patent. Copyright. Sued by Apple. Then by Google. – kenny Mar 14 '12 at 22:33
this is the most complicated and may be the best way we found, we are trying to solve this problem much easier if we can ;) – bahadir arslan Mar 15 '12 at 06:56
@bahadirarslan Well, if this is a software-only problem (i.e. you can't attach any accessories to the phones or locations), I just suggested this *may* be less complicated than trying to create a new, non-GPS positioning service from scratch. I agree that my suggestion would be very difficult to implement, but so far it sounds like the easiest of the options offered here. I wouldn't be at all surprised if someone came up with something easier, (in fact, I hope someone does) but Olin's answer seems to indicate that you have to start from scratch and compete with other experimental solutions. – Mutant Platypus Mar 16 '12 at 13:57
@bahadirarslan If you eventually do go with a software-filter solution, I would point you to [Stack Overflow](http://stackoverflow.com/) or [CSTheory](http://cstheory.stackexchange.com/). People there will know more about exactly how difficult it may be to implement and how much of a load it will place on the phone's processor. – Mutant Platypus Mar 16 '12 at 14:02

score 3 · Accepted Answer · edited Mar 20 '12 at 08:43

3

Check out Qubulus. Their indoor positioning API uses an RF fingerprinting method for Android devices. Right now, your software-only options are limited for iOS. Best tip there is to wait for iPhone 5. I'm not familiar with options for Blackberry.

A full list of indoor positioning technology providers is here: http://bit.ly/indoornavigation

You might also gain some insight from reading "Why Indoor Navigation is So Hard". The O'Reilly article mentions the Cisco MSE option, which is likely a better fit for your needs than a GPS repeater.

edited Mar 20 '12 at 08:43

clabacchio

13,481
4
40
80

answered Mar 20 '12 at 04:08

nicksuch

146
3

thanks, i will check this too. At weekend i found Navizon ITS. Navizon has a product that allows you to track wifi enabled devices and it's API allows you to find routes. – bahadir arslan Mar 20 '12 at 09:16
1

I am accepting this reply as answer because this file: http://bit.ly/indoornavigation . Thanks for everyone who helped me.. – bahadir arslan Mar 21 '12 at 08:46

score 1 · Answer 4 · answered Mar 14 '12 at 21:18

iPhone uses GPS, then CDMA, then WiFi geolocation to provide the phone's location. The phone API provides hooks to the current location which should be somewhat usable, even indoors. If this not accurate enough, the solutions are experimental and expensive.

This paper is a really good overview. http://www.sis.pitt.edu/~dtipper/2011/Survey1.pdf

score 1 · Answer 5 · answered Mar 14 '12 at 23:58

1

I have looked at this problem before. Apart from what is being said I can only suggest the following. You can get a fake wifi ap using a client chip and build a sub 5$ AP that you can use its mac id. you place these in your target area and as you navigate you look for available AP s and compare their mac to a database. I calculated that you can reach accuracy Of 15 20 meters.

answered Mar 14 '12 at 23:58

Ktc

2,226
3
26
48

this solution sounds acceptable, i will share this with my team mates. – bahadir arslan Mar 15 '12 at 06:59
Bahadir, kolay gelsin.. – Ktc Mar 15 '12 at 09:00

score 1 · Answer 6 · answered Mar 15 '12 at 08:40

For a similar problem, but may work also in this case, I have used AHRS (Attitude and Heading Reference System) using a Kalman filter.

Basically what you need is:

An absolute positioning system, like signal strength from some wireless spots or similar; this has not to be very accurate, but gives the reference for the rest of the system;
An accurate way to measure movements, like an IMU (Inertial Measurement Unit), with accelerometers, gyroscopes and magnetometers; this will give you a good estimation of the movement, and at least the accelerometers are provided by most of the smartphones.

With Kalman filter, you can use the first block to make a prediction of the position, and then the second to apply the movement to that prediction. This allows you to relate more measurements of the absolute position and have the convergence of the solution to a very accurate result.

Note that if you have a good starting spot (e.g. a checkpoint with - say - an NFC tag) the precision can be improved a lot.

You may also try with particle filters, where movement is used in combination with measurements in a similar way, to obtain the convergence of the position to an unique spot.

This sounds like a banner application of particle filtering, as you're bound to get multimodal location probability density functions. — drxzcl, Mar 15 '12 at 10:24

Indoor Location Positioning

6 Answers6

Linked