# channel model ## Simple Expression $y(t)=h(t)*x(t)+n(t)$ ^[* denotes convolution] $n(t) = AWGN\ (noise)$ This captures the essential characteristics of the system and we've performed some analysis that provides some insight about the system. ## Why don't we use Maxwell's equations? What do we need for Maxwell's equations? Everything! Refraction index, geometry, etc. It's unlikely that you can get this and you want it to work in any environment. We don't know much about the channel. It's also computationally demanding. ## Dimensions of the channel  It's completely random in all dimensions! Usually model by a random variable. If the time duration is very short, you can treat it as a constant (ns), in that moment the channel might be constant. This depends on the user characteristics. Things are relatively - it might be fast to your eyes but not compared to other signals. ## Typical Approach 1. Field measurements for a class of environment (e.g. indoor) 2. Model fitting ## 3-level Model (statistical) [Large scale propagation](large-scale%20propagation%20models.md) [small scale propagation models](small-scale%20propagation%20models.md) ### Joint effects  Averaging over different distances gives the levels of modelling. # References 1. [TELE4652-lecture-03](../../../Spaces/University/TELE4652/Lectures/TELE4652-lecture-03.pdf) # Footnotes