WiFi signal strength distibution

I wonder how wrong is it to assume that in general case for a non-directional antenna the wifi signal level distirubiton can be described with the function $$\y(x) = b - \frac{k}{(x+a)^2}\$$, where x is the distance module from the point of measuring the signal lvl to the position of the antenna (wifi-router or other signal source). The parameters $$\ b, k \$$ and $$\a\$$ (all positive) can be determined using 3 different points (one of them in the closest proximity to the source antenna).

I need all this for my graduate project in which I need to create a model which would provide the wifi signal level in desired points in a room after taking some measurements.

• Valid pretty much only in open space or an area (or volume) surrounded entirely in a material that absords all radiation at WIFI frequencies. Any reflection will cause "hot spots" where the signal is stronger than your (simple) model predicts.
– JRE
Apr 21, 2019 at 18:19
• @JRE Well then my model would be able to predict the "worst case scenario". I think I can work something out if critically needed later if I'm right about my initial question. Apr 21, 2019 at 18:47
• Your model won't predict "worst case." It will predict "simplest case" which doesn't apply in a space with reflections and obstacles.
– JRE
Apr 21, 2019 at 19:23
• @JRE I have an idea on how to predict obstacles. Reflections are a bit harder though. Apr 21, 2019 at 19:28
• If you have an unmodulated sine wave, then reflections can cause hot spots (areas of constructive interference) and cool spots (destructive interference). But if your sine wave is modulated, and is conveying data, then, in addition to simple sine-wave interference, you also have interference between subsequent data sent through the channel. Apr 21, 2019 at 20:04

I wonder how wrong is it to assume that in general case for a non-directional antenna the wifi signal level distirubiton can be described with the function quadratic loss

Pretty wrong.

WiFi is an indoors technology, and your model would apply to free-space only.

You're not taking into account any small-scale fading. That won't do.

I need all this for my graduate project

Then you'll need to do quite a bit more research.

in which I need to create a model which would provide the wifi signal level in desired points in a room after taking some measurements.

What you want is typically done with ray-tracers and/or finite element simulations.

• >WiFi is an indoors technology, and your model would apply to free-space only. I use the equation to figure out the "clear view" signal level and then account for obsticles which should reduce the signal level on a constant value specific for the obstacle. Still don't know how to deal with reflections, noise sources and interfiering sources. Apr 21, 2019 at 18:42
• Thanks a lot for your answer, I will look more into it. I would not be able to completely redo my work considering the time limit. You've still provided the answer I was looking for. I think the biggest problem that measuring the signal level with smartphone gives only a scalar value which already accounts for all kinds of losses and interferences. It would be much more easier if there were fields vectors. Apr 21, 2019 at 18:57
• a smartphone signal quality meter is in no way appropriate for a graduate project, sorry. It gives you an RSSI estimate, which often is more dominated by interference than by actual signal strength. Also, no, your "clear view" baseline model simply does not apply. There's not only shadowing, but also things like constructive interference. It really doesn't work like you're trying to potrait it, and if your advisor knows anything about radio propagation, trying to solve it like that will not be good for your grade. Apr 21, 2019 at 21:04
• I'm a software engineer student making a mobile app, so there is not much else I can change, just trying to put what I have to the best use possible. The thought process was that this approach would give me a good-enough quality and precision for a mobile app. Apr 22, 2019 at 9:05
• sorry, not the case. Apr 22, 2019 at 14:12