Receiver BER simulation?

[bardo84]

Could Causal.jl do a bit-error simulation for a receiver chain? Impairments are thermal and other noise, receiver nonlinearities and multipath reception. The usual way is to repeat a telegram reception many times with different generated noise. Thanks for your work!

[zekeriyasari]

I could not get all the details (receiver chain, thermal noise, receiver nonlinearities, etc.), What I understand from here

Could Causal.jl do a bit-error simulation for a receiver chain? Impairments are thermal and other noise, receiver nonlinearities, and multipath reception. The usual way is to repeat a telegram reception many times with different generated noise.

is that there exists a chain of systems (e.g. nonlinear dynamical systems whose dynamics evolve by SDEs etc.) in your system. Causal.jl provides you with all these different kinds of components. You may refer to the documentation for the details of the components.

Or, you can take a look at my other repo here. if you are interested in a BER simulation in a classical digital communication system.

[bardo84]

Hi @zekeriyasari , thanks for replying.

I could not get all the details

Interesting schism I could observe many times: the digital folks have no idea where the impairments come from, the analog folks (me included) have a good view of where noise and nonlinearities arise, but lack the higher level view when it comes to noise folding, side bands and how it translates to BER :-)

Thanks a lot for pointing me to your package, was not aware of it.

nonlinear dynamical systems whose dynamics evolve by SDEs

Apart from switching/mixing, a receiver is LTI. The system does not evolve, but for some components tolerances are known. So, actually I was asking if classical BER calculation for systems with tolerances can be sped up by any fancy math like SDE or other trick in MTK.

[zekeriyasari]

Hi @bardo84 . Thank you for your explanations.

Apart from switching/mixing, a receiver is LTI.

I agree with you here.

The system does not evolve

I got confused here, again :) Just to be sure, below is shown the structure of a matched filter receiver (From Proakis and Salehi's Digital Communications book). The matched filters are LTI systems representable by state-space equations, which are ODEs. So, I think, the system does evolve.

but for some components, tolerances are known.

This sounds like you are looking for a tolerance analysis tool based on something like interval algebra. But Causal.jl has nothing to do with interval algebra.

So, actually I was asking if classical BER calculation for systems with tolerances can be sped up by any fancy math like SDE or other trick in MTK.

Frankly, I do not know how SDE's are used for such a purpose, unfortunately.

[bardo84]

My confusion, the states and outputs evolve of course. Topology and parameters don't. Partly answered here. Thx