Disclaimer - I am in no way commercially linked to any RTLS company. References are for information only.
We have all seen sci-fi movies where our hero is being monitored in 3d as he moves his way round a building. As is often the way in the movies, the reality is much less glamorous.
RTLS over WiFi uses tags attached to devices or people, and then a high density deployment of access points to monitor the tag and best-guess as to where the tag (and therefore the equipment or person to which it's attached) is located. The two current market leader for this technology in the WiFi space are Ekahau and Aeroscout.
Here is a picture of some Ekahau tags for scale, I placed them next to a pen and a Post-It. The flat one can be tied to a device, the stubby one can be worn round a wrist.
There will be an overlay locationing system (various vendors) tied in with the wifi that is able to use metrics between the tags and the access point to guess where the tag is, and position it on the map.
The WiFi needs to be very carefully designed. Design Rules of thumb when we plan networks:
* WiFi for data only - must see one access point with a signal strength of -67 db
* VoIP - must see two access points with a minimum strnngth of -67dbm
* Location tracking - must see three access points with a minimum strength of -67dbm.
If you don't plan correctly, you face a significant problems. The system will not be able to accurately place the devices and may even locate them to the wrong floor in multi-floor environments. With careful planning, it can be very accurate - the best I have seen is accurate to within 15ft and always the right floor - Tallahassee Healthcare, Florida.
So the challenge is to squeeze a lot of access points into a small space. Here is a design for a single floor measuring 110 ft by 70 ft - this space contains 22 Cisco access points on 2.4GHz, so three channels are in use.
In order to reduce any co-channel interference (access points able to see each other on the same channel), the access point power was reduced to the bare minimum and extremely detailed attention paid to the use of directional antennas.
Here is the map showing the TOTAL interference, the green is 0% and yellow is 10%
This is a truly remarkable achievement, and takes great skill, time and dedication.
So why go to all this trouble ? Well, most of my experiences come from healthcare (hospitals), so I will give you some examples of the use cases there, though you may be able to think how this may help in education:
Crash carts for heart attack patients are very expensive so each floor may only have one or two. If you need on quickly, RTLS will show you on a floor plan where it is.
Infusion pumps are expensive to buy so one hospital leased a multitude of them, tagged them then tracked them as they were used. They worked out how many they really did need, and how much use they got, thereby working out if it was cheaper in the long run to continue to lease them or buy them. This use case saved over $10 million alone.
Valuable equipment can be "geofenced", so if it is moved outside a boundary an alarm will sound as a means to prevent theft.
People can be tracked - key staff can be located, patient movements can be tracked and analysed, and so on.
There are many other uses too, centred around the system giving you information and you using that information to make business critical decisions.
I hope this helps answer questions you may have.
A final word - you may all have seen the map functions in Ruckus, Aerohive, Aruba Airwave, Cisco and others that shows client devices on a map. This can't really show you where the device is, the placement is just random - what the system is really trying to do is show you the devices and the access point to which it's associated, so it will add an icon to the map but it has no way of showing the true location, and doesn't even try to guess.
Last edited by neilmac; 9th October 2013 at 11:03 AM. Reason: typos
How much is a typical tag? Is it practical to buy cheap bulk ones and install an open source RTLS system, or make our own?
When a device receives a frame, it goes through a process in order to decide what to do with it. It's quite complex, but I will try and explain as it will show why your suggestion wont work;
1) Detect frame energy, listen for the frame header info and read FRAME_LENGTH field, set clear channel assessment>NEAGTIVE for this value.
2) Attempt to read frame header and read BSSID address (if I am an access point, is it me?) or DESTINATION_ADDRESS (is it me ?)
3) If it's me, prepare to reply (ACK)
4) If it's not me, read DURATION_VALUE and set NAV timer to duration value, add 0- CwMin slot times and wait till time expended. DROP FRAME - so, any valuable information that the frame would contain in order to location track is dropped at this point.
So, unless you are in promiscuous mode, you drop the frame. This is how frame capture works when analyzing networks - you tell the card driver not to drop the frame, but in order to do that you become passive, listening only. The access points would need to be specially tweaked in order to pass on the metrics, something that isn't inherently built in.
That's why tags are used - to derive and deliver metrics to the overlay system.
Furthermore, unless you have a high density of access points, the system would not have enough information to accurately pinpoint you.
I have never heard of an open source RTLS with home made tags, it would be an intriguing project - I see great things being done in schools with Raspberry Pi, maybe someone could look into it as a project (I may even be able to drum up some sponsorship if it's looked at seriously).
As for cost, I have no idea, you would need to approach the vendor or a reseller to get an idea of pricing.
Update - I found this open source rtls, very interesting - if you investigate it let me know how it looks: http://sourceforge.net/apps/mediawiki/grailrtls/index.php?title=Main_Page
Last edited by neilmac; 9th October 2013 at 12:15 PM.
I'm not so sure, I am pretty tight with a few of the senior people at Ruckus, so I can try and find out more.
Ruckus recently acquired YFind:
It's all about Location | Ruckus Wireless
Ruckus starts tracking indoor location with YFind acquisition — Tech News and Analysis
My guess is it's going to be used for user analytics within generic areas but not really pinpoint accurate. Think "Minority Report" where you will be pushed advertising relevant to where you are in a shopping mall - based on where you are associated, probably.
If could be the inbound gain and the Beamflex antenna array will provide some proprietary features to help plus YFind is undoubtedly going to be an overlay. The challenges are the same - you need more access points than normal to be accurate, and the inherent behaviour of the devices doesn't lend itself to locationing.
It's definitely a growth space, I will see if I can get any technical specs.
Last edited by neilmac; 9th October 2013 at 03:57 PM.
Further info -
I wasn't sure if you knew the ZoneDirector settings are optimised for working with Ekahau tags:
I also read that they are planning to have systems to allow locationing of school assists such as tablets, so you are right that it will probably be a controller-based tagless solution.
All the more reason for having the AP placement professionally planned, and for the budget to support it.
Last edited by neilmac; 9th October 2013 at 04:11 PM.
Well, this is turning into a very interesting post.
It's true that Ruckus do have something in the pipeline. I spoke with a couple of high up friends at Ruckus. They couldn't divulge too much, and I can't repeat what was divulged, but it's definitely on the way, and I am given to believe that a standard install would support it without having to install additional access points, or very few additional access points.
Sorry, I have no idea about that.
There are currently 1 users browsing this thread. (0 members and 1 guests)