When I am thinking about communication channels, like wireless ones, how is the statistical overview and modeling of the channel obtained? Mainly we are dealing with all random manners here? How researchers could expect the received signal and give it a mathematical model? I have a problem about understanding the gap between research and practical.
You are right that we cannot completely model a wireless channel with precise accuracy. however we can predict the behaviour of the channel through repeated experiments and then propose a probable model. It is similar to the electron cloud of an atom where we say that there is 90% probability an eelectron lies within such region.
Thanks for that , but how are the models we have these days in the books were derived? in experiment only? interesting
Thank you for your answer
According to my experience a measurement campaign is done first and the received signal is stored. Afterwards a post processing is done. One tries then several fading models and sees which is the one which fits best the observation (measurent).
You can find below a publication where you find some details about the development of a mobile satellite channel model. I think for terrestrial models a similar process is used.
https://www.researchgate.net/publication/224154028_Mobile_satellite_channel_with_angle_diversity_The_MiLADY_project?ev=prf_pub
Conference Paper Mobile Satellite Channel with Angle Diversity: the MiLADY Project
There are a number of books describing wireless channel attenuation modelling. My favourite is Rappaport's "Wireless Communications Principles and Practice" from 2002.
See chapters 3 and 4.
from the theory of Statistics, we can devise a model and then if the real field data closely conform to that model we say we have modelled the wireless channel. Else we look for other models. That is why we have different models for same attribute calculation like (Log-distance path loss model, log-normal Shadowing path loss model). Similarly for signal prediction we have Okumura model, Hata model,PCS , Walfisch and Bertoni model etc.
Yes, Rappaport's "Wireless Communications Principles and Practice" is an excellent book.
Through repeated experiments conducted under different physical conditions and also by using the laws of statistics and probability theory we can make the most approximate model which in most of the cases behaves as expected. This way the mathematical model of the practical situation is made.
The best example is the additive white gaussian noise which was modeled with the help of gauss distribution.
Dear Mousa Al-Aomari,
Take help from the course "Signal and Systems" where one gets comprehensive mathematical modeling of signals and systems. As mentioned in most of the posts that define parameters, measure and apply staististical model to get the model of the channel.
The best way is prediction but this prediction is not totally random. Most of the models presented in literature are derived from experiments. Most popular measurement campaign is done by sending signals which are close to impulses on the channel. A receiver then measures the response of the impulse by the channel. This is done repeatedly from various transmitter positions and velocity. After measuring all these responses an empirical mathematical expression or spline is fitted. Again this expression is not universal. Even currently work is going on fro accurate modeling of wireless channels.
nice summary by Ismat specially a hint of research work for millimeter waves. Yes we need new methods for channel estimation and derivation of channel models.
Mr. Ismat can you share the reference for the work in mm waves by rappaort at NYU.
- Badges
- Science topic