WiFi radio planning for short distances

[rooftile]

Hi all

I'm trying to develop an application in 802.11 field, short distances. The scenario consists of WiFi end points at different distances and the master point must be able to detect which point is the closest one. The distances are about 10 meters and can reach a maximum of 100 m. The antennas I'm going to use have a gain of 11-13 dB.

Theoretically I know the dependence between the received power and the distance. In reality, is it possible to distinguish a station at 20m from another one at 40m in outdoor environment?

Some hints about the chipset I can use to develop the end points (I think that I should take care of the resolution in detecting the received power from the remote end points)?

Thank you

 

[bigdogguru]

GainSpan makes a relatively inexpensive WiFi module. I've seen them for less $25 at quantity.

GainSpan produces several variations, this is one of their less expensive modules:

GainSpan Module

I've had good success with them in the past.

BigDog

 

[rooftile]

Hi bigdogguru

Thank you very much for your fast and useful reply!
Do you think that with these modules I'm able to distinguish two nodes at 10 meters and at 25 meters, as an example? Or you think that this is independent from the module?

Rooftile

 

[bigdogguru]

Do you think that with these modules I'm able to distinguish two nodes at 10 meters and at 25 meters, as an example? Or you think that this is independent from the module?

Distinguish? I guess I do not understand the question. Do you mean will they be able to successfully link with the WiFi access point?

If so, 10 meters yes, 25 meters depends on the environment. If there are numerous rooms with metal studs it might be a challenge, if it is an open area for the most part, you shouldn't have a problem. You could use directional high gain antennas to improve the situation.

BigDog

---------- Post added at 10:49 ---------- Previous post was at 10:46 ----------

I just noticed you originally mentioned using a high gain antenna so you shouldn't have a problem.

 

[rooftile]

The environment consists of different nodes at different distances (fixed, known). And I would like to know which one is the closest from the master node. Could the measurement of the received power a good parameter to do that? It's a sort of neighbours discovering, but I'm interested just at the nearest node.

I'm using a directional antenna with gain 11-13 dB.

R.

 

[E Kafeman]

Signal level as distance indicator, may work, may not. It depends on actual scenario and what error rate that can be accepted.
In an ideal free space do it work without problem but in most normal environments can local reflections/shadows cause signal variations of 10 dB or more. It is then not possible to judge which node that is 20 or 40 meters away. If it is a mobile/handheld units do that also that increase complexity as signal level will depend on how the TX unit is hold or placed relative the body.
Some of these problems can be reduced for specific environments with aid of smartness at receiver end.
By analyzing received signal level over time for nodes that moves around or if environment changes in some other way, can problems due to temporary variations be reduced, but not eliminated, for example.

 

[rooftile]

Hello Kafeman

With "what error rate that can be accepted" you mean the sensitivity of the receiver or the resolution? or both?
Could it help if I use the same modules as TX and the same power (set so that the maximum reachable distance is 30-40 meters)?

R.

 

[E Kafeman]

Most locations will have far from ideal (free space) coverage. That will cause a difference in measured signal level compared to the ideal situation.
Your local situation can be analyzed, max RF field variation, that not can be compensated for will be an error factor.
I assume same TX power and antennas from all units. If not, that must that be compensated for individually at RX station or also added as an error factor.
Add/multiply all errors to find out total error rate. Use this rate to calculate likeness that your receiver can find correct TX.

 

[bigdogguru]

The environment consists of different nodes at different distances (fixed, known). And I would like to know which one is the closest from the master node. Could the measurement of the received power a good parameter to do that? It's a sort of neighbours discovering, but I'm interested just at the nearest node.

If each WiFi module is fixed and at a known distance, why not assign each module a static IP and reference them accordingly.

BigDog

 

[rooftile]

If each WiFi module is fixed and at a known distance, why not assign each module a static IP and reference them accordingly.

BigDog

Because the nodes move normally. They remain fixed just when they start the communication with the master node. I have just to realize the nearest one.

---------- Post added at 12:02 ---------- Previous post was at 11:56 ----------

Most locations will have far from ideal (free space) coverage. That will cause a difference in measured signal level compared to the ideal situation.
Your local situation can be analyzed, max RF field variation, that not can be compensated for will be an error factor.
I assume same TX power and antennas from all units. If not, that must that be compensated for individually at RX station or also added as an error factor.
Add/multiply all errors to find out total error rate. Use this rate to calculate likeness that your receiver can find correct TX.

I confirm: same TX power and antennas. I will try to think about the received power and the likeness and the ... cross the fingers!

Are there algorithms/implementation that do already this evaluation for WiFi nodes?

 

[E Kafeman]

Are there algorithms/implementation that do already this evaluation for WiFi nodes?

WiFi do not differ from other radio sources. The problem is to set up a model for your location. Outdoor exist several standard models and analysis tools. Okumura-Hata can be a good start.
Indoor can be much more complicated as almost every location is very unique.
Practical Modeling and Prediction of Radio Coverage of Indoor Sensor Networks
**broken link removed**

 

[bigdogguru]

Hi R,

I found this document while researching another topic:



There are some interesting algorithms presented, maybe one will fit your purpose.

BigDog

---------- Post added at 15:46 ---------- Previous post was at 15:42 ----------

Another interesting document:

**broken link removed**