Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
They are usually random for wireless channels and follows a known pdf. You can directly use MATLAB's Rayleighchan function or search the web for some famous channel profiles (fixed value bins)
your code is not wrong in syntax but the channel you generate is not practical. do you have any syntax errors in this code ? The delay spread is the length of the channel (16 * sampling period) and the symbol time is the length of the symbol (usually = sampling period), the degree of ISI can be quantified as delay spread/ Symbol duration = it means that every symbol will interfere with how many adjacent symbols (in your case 16).
Use Rayleighchan function in MATLAB, it specifies the channel taps, the time delays, the doppler shift and the sampling frequency. Check rayleighchan in MATLAB help for further information
your code is not wrong in syntax but the channel you generate is not practical. do you have any syntax errors in this code ? The delay spread is the length of the channel (16 * sampling period) and the symbol time is the length of the symbol (usually = sampling period), the degree of ISI can be quantified as delay spread/ Symbol duration = it means that every symbol will interfere with how many adjacent symbols (in your case 16).
The channel length is measured in symbol time, and I believe there is a difference between sampling period and symbol time. Suppose that the sampling frequency is 8 Khz, i.e.: we have 8 ksamples/s, i.e.: sample period is 0.125 ms, and each sample is represented by 8 bits, then we have 64 kbps, i.e.: the bit duration is 15.625 microsec. Now suppose we use 8 PSK, i.e.: 3 bits/symbol, then the symbol time will be 46.875 microsec. So, the symbol time is not the same as the sampling time.
There was a syntax error, but I fixed it when write:
d=2.*x-1;
conv(d,h);
Again, why is my code not practical? and how to make it so?
Of course symbol time and sampling period are not always equal, I meant that usually we use Nyquist pulses so we need only one sample per symbol to describe it. As for your channel, you can simply use the built in rayleighchan function. The channel you're using have random path gains that are gaussian and this is not encountered in nature.
But you know that a Rayleigh fading channel describes only the amplitude of the channel. But the channel coefficient itself is Gaussian with zero mean.
[/COLOR]Use Rayleighchan function in MATLAB, it specifies the channel taps, the time delays, the doppler shift and the sampling frequency. Check rayleighchan in MATLAB help for further information
well, i used "rayleighchan" before but because i depend on simulation from the book "simulation and software for radio channel" it did not work, because the aurthor generate a rayleigh channel with different way. so it didnot work. and he didi not explain his way.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
