Thanks for your answer Dr. Emil.

Do you also think I could adopt the model shown below, where the large-scale fading term is not associated with each one of the paths?

z[n] = g[n] x (h_m[n] * x[n]), (1)

where, "x" and "*" denotes multiplication and convolution respectively, g[n] models the large-scale fading and the second term, (h_m[n] * x[n]), corresponds to the distortion introduced by small-scale fading.

Dear Felipe,

According to the model you brought in the question, the large scale fading is represented by the gain factor g[n]. While small scale fading is represented by the coefficients hm[n]. So, your interpretation is okay.

I think it is also the the most compact way to represent the multipath effect, where one observes the resultant summation of all signals coming from the different paths and then model the channel as an effective large scale fading part representing the average path loss at that point and small fading varying statistically with time. This model with its coefficients may be satisfactory to represent the variation of the signal at a targeted location in the channel.

Best wishes

Yes, it makes sense to assign different large-scale fading values for the line-of-sight path and the non-line-of-sight paths.

Thanks for your answer Dr. Emil.

Do you also think I could adopt the model shown below, where the large-scale fading term is not associated with each one of the paths?

z[n] = g[n] x (h_m[n] * x[n]), (1)

where, "x" and "*" denotes multiplication and convolution respectively, g[n] models the large-scale fading and the second term, (h_m[n] * x[n]), corresponds to the distortion introduced by small-scale fading.