Slashdot Mirror
Mapping a Wi-Fi Network?
NivekEnterprises asks: "At my school where I am an Electrical Engineering major A Wi-Fi network has been set up, and is running in several buildings. Since the coverage extends outside several of my friends and I are trying to map the signal strength. Basically we are going to take a map of campus and walk around with a Wi-Fi enabled laptop, marking as best we can. Is there an easier way to do this? Has anyone else done something similar and is willing to share some their insights?"
Belkin (I believe) makes a Wi-Fi hotspot detector.. coupled with a GPS receiver and GIS software I think you could do a pretty good job of it..
serveral times... over and over and over.
Consult google:
google
The program will show you what spots are hot, what is not... Good luck!
I think one or two people may have done something similar... Maybe Slashdot might be of assistance
Why can't I moderate something "Wrong" or at least "Grossly Misinformed"?
Why are you making this map? Anyway walking around with your laptop computer may be the cheapest way to get your map. You can always just ask the school for where they placed all their wi-fi equipment. It should be public information and you can largely base your maps on that.
Get yourself a GPS usb device. You can collect raw data from it and easily overlay the signal strength to a school map. You might run into trouble though if the map isn't scaled properly though. This way you can just walk around and have the laptop do everything.
There are several methods by which this could be accomplished.
The first is to create a link budget using the two-ray approximation (1/R^4 attenuation) and the estimated antenna patterns. This allows one to bound the maximum range as a function of antenna orientation and receiver sensitivity. Most likely this is the analysis that was done when installing the network.
A second and much more enjoyable way is to use a ray-trace simulation program such as Wireless InSite to model your campus. This model will pick up multipath effects and folliage losses.
The most time consuming but most accurate method is to walk around with an antenna, measuring the power as one goes. It should be noted that when one measures in a given location the power will change over time sometimes quickly. Known as Rayleigh fading, it is due to time-varying multipath from a dynamic environment.
I recommend tracking down a communications professor in EE and borowing their copy of Wireless InSite. If you pitch it right you could even get a credit of independent study from it.
Michael.
Linux : Mac
Get 2 more wi-fi laptops and triangulate.
If you use Linux on your laptop, kismet will interface with GPS devices, and do signal power interpolation to find signal sources. It will also mark everything on a user-supplied map. Good Luck.
www.kismetwireless.com search sourceforge for gpsd. Hook your GPS to your linux machine and get gpsd running. Configure kismet (sniffer) to use gpsd. Kismet will record the gps location, SSID, and strength in a log file. There is even another tool included to automatically generate maps for you. --Chris
Post the location and the SSID outside of the building and on the newsgroups. The warchalkers will do the rest for you. :)
Forget the map. What you need to do is find a way to get an analog signal strength reading (either with a device designed for the job, or hacking something on a laptop to (say) pipe a running digital signal strength value out the dsp to the earphone jack) and run it into an op-amp setup that will 1) give you decent current and 2) give you an inverted signal as well. Run these two signals to two actuators. Hook the inverted one to a can of red spray paint,and the un-inverted one to a can of blue spray pant.
Do this with the power off.
Go outside, make sure the cans are facing away from you, power up, and walk around.
From then on, the parts of campus with a blue tinge will be the ones with good signal, and the red areas will be the ones to avoid. Easy rule to learn, and no maps to deal with.
-- MarkusQ
P.S. You may need to go over some parts of campus (esp. plants, artwork, etc. that may be strongly coloured) with a coat of gray primer first. I suggest you just play around with it till you get something usable.
A Finnish company called Ekahau has a product called Site Survey that does map a wireless network visually and quite nicely, I might add.
We're currently planning a Wi-Fi network for our school as a final project and Site Survey seems to be a nice tool for figuring out the best places to set up APs so we get maximum coverage.
The Site Survey product can be found hereToo bad it's windows only...AND costs a bundle ( starting at $1995 afaik )
I work at a university that's done this.
We tried a few different approaches. It turned out that the fastest approach was to send someone around with wireless card tools that showed SNR (signal to noise ration). In offices, the SNR at the most likely use location (e.g. the desk) for the "best" AP (access point) was measured. In classrooms, they broke big rooms into smaller chunks, then found the best SNR for each chunk.
An approach using GPS was tried, but it took longer to wait for the GPS to stabalize than it did to enter the data by hand.
An approach using pedometers was also tried, but again, it took longer to enter meta-info than it did to just enter point measurements.
--
Build a small blimp with a payload consisting of an embedded linux SBC with GPS and WiFi recievers. Program it to fly along arbitrary signal strength boundaries (think lines of flux) and chart its location either to an onboard storage location, like CF, or broadcast it back to a server on the WiFi network, then layer the GPS info onto your campus map like its GIS.
Who want's to go out doors and walk around exposed under that burning ball of gas? Really?
Well, you can still do that, but here's what I would do before going outdoors, and this may even provide the level of accuracy you need.
First, make a digital map of your school. As simple as scanning/cleaning up a paper map given to new students, or snag one from your local city hall's zoning department (of course, doing that may end you up on an FBI watch list).
If you really want to geek out, grab a friend and have them do a rough 3D model of the school.
Second, plot the WiFi APs on your digital map.
Third, look up the specifications for each model of AP being used. You should be able to find these in the product documentation/packaging or just go online to the mfr's website.
Fourth, plot circles around the APs on your digital map according to the range specs. I would plot both the unobstructed and obstructed ranges (most WAPs I've looked at the specs for provide clear line of sight and obstructed figures).
Also, I would use a thicker line for the circle that is more applicable to the AP's location (or different opacities on spheres if you're doing this in 3D).
Finally, if that's not accurate enough, or if you're one of those freaks that likes sunlight, go for a walk with a laptop/pda, a copy of your digital map and your notation device of choice (paper, text editor, spreadsheet, etc).
Choose a granularity for your markings (say 10%).
Now, go to each AP and start walking in a straight line until your signal strength goes down by 10%, write down the distance. Repeat.
It doesn't have to be a perfectly straight line, but just make sure you can measure your distance accurately.
If there are significantly different amounts of obstructions near an AP, walk a few lines through the extreams.
Now, take your data and plot new circles from the APs, cutting/bending them as needed for where you walked multiple lines for varying obstruction levels.
Now you should have a nice, topographic looking map of your WiFi network.
DONT PANIC
GPS is the best way to go. Failing that, I used a system a bit like the following:
Every ~25 seconds, laptop records WiFi signal strengths and network names and such to record #N, and emits a pair of tones corresponding to the last digit of N (I used musical intervals and the first tone was always the same; for the zero digit, the second note of the pair was null---in other words, a single tone). I carried a clipboard, which I would number as I went along, and fill in my approximate location. I could make sure I kept sync between laptop and clipboard based on the tones. I also had the laptop emit, 1.5 seconds later, another pair of tones, an octave up from before, with a similar system, telling me how many network APs were nearby; then I could know if I was in a "hot" zone or a "dead" zone and continue my explorations in light of that information
Yes, I used headphones. Yes, people looked at me funny... but oh well.
I wrote a bunch of Scheme code for The GIMP that plotted all the records on a multilayered campus map (I would have to figure out my coordinates from my clipboard notes by hand, later, but it wasn't hard, and once I got into the swing of it, it went very quickly), then plotted what kind of speed was available at each access point, and tried to estimate the position of any given AP node based on where it could be detected from.
In retrospect, two things I would have done differently:
* used GPS. would have saved a LOT of work. but I'm poor, and I don't yet have a receiver...
* used the signal strength to weight the AP-locating algorithm, so it would be more accurate
* made an additional layer, showing the signal strength using color shading or something
--TheOrangeSquid Is it any wonder things seem so awry? We swim in a sea of confusion and don't have to think to survive
"Can you hear me now?"
1) make a clickable map on a campus web-server
...
2) have helpfull students point there web-browsers
homepage to that web-server.
3) have helpfull student look a WI-FI signal
strength.
4) have helpful student click on map for there
present location
5) have web-server present student with a box
for choosing singal strength
6) have web server display message
"thank you for your help".
7)
8) Profit!
it's "Can you ping me now?"
Isn't it interesting how you come to recognize posters based solely on their sigs???
Slightly off-topic, but what tools could be used for the inverse process. That is, I want to determine the best channel to select for a particular environment based on noise level readings (from other wi-fi APs, microwaves, cordless phones, etc). Is there an app to run on my PC and use my wi-fi card as a spectrum analyzer?
I highly recommend KisMAC, works great, never played with the GPS bit but it autogenerates a map. Two cavets - it only works on Jaguar and, while it supports Airport Extreme, no passive mode. It does, however, support a long list of PCMCIA cards and quite a few in passive.
r en /kismac/
http://www.binaervarianz.de/projekte/programmie