Hi, I have a general about the usage of equalization for non-coherent systems. I'm considering Differential PSK for a single carrier SISO system. I consider different channels with different no., of taps. But I'm not sure what type of equalizer should be used for a such a system. Thanks. _____________________________ Posted through www.DSPRelated.com
Equalization in non-coherent systems
Started by ●March 26, 2014
Reply by ●March 26, 20142014-03-26
On Tue, 25 Mar 2014 22:17:03 -0500, "Dan25" <98432@dsprelated> wrote:>Hi, > >I have a general about the usage of equalization for non-coherent systems. >I'm considering Differential PSK for a single carrier SISO system. I >consider different channels with different no., of taps. But I'm not sure >what type of equalizer should be used for a such a system. > >Thanks. > >_____________________________ >Posted through www.DSPRelated.comThe easiest thing to do is use a Constant Modulus Algorithm. There are downsides to CMA, but it works reasonably well at high SNR. Otherwise it gets very tricky. With some special tricks one can get reasonably close to coherent system performance, even at lower SNRs. This is an implementation of my favorite method: http://www.anchorhill.com/LinkedFiles/AH1000_MPRE_datasheet_v1.0.pdf Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
Reply by ●March 26, 20142014-03-26
>On Tue, 25 Mar 2014 22:17:03 -0500, "Dan25" <98432@dsprelated> wrote: > >>Hi, >> >>I have a general about the usage of equalization for non-coherentsystems.>>I'm considering Differential PSK for a single carrier SISO system. I >>consider different channels with different no., of taps. But I'm notsure>>what type of equalizer should be used for a such a system. >> >>Thanks. >> >>_____________________________ >>Posted through www.DSPRelated.com > >The easiest thing to do is use a Constant Modulus Algorithm. There >are downsides to CMA, but it works reasonably well at high SNR. > >Otherwise it gets very tricky. With some special tricks one can get >reasonably close to coherent system performance, even at lower SNRs. > >This is an implementation of my favorite method: > >http://www.anchorhill.com/LinkedFiles/AH1000_MPRE_datasheet_v1.0.pdf > > >Eric Jacobsen >Anchor Hill Communications >http://www.anchorhill.com >Hi Eric, Thanks very much, Is there any good reference to implement the CMA algorithm ? Are there any algorithms which operate using Channel State Information ? Best regards. _____________________________ Posted through www.DSPRelated.com
Reply by ●March 26, 20142014-03-26
On Wed, 26 Mar 2014 02:05:58 -0500, "Dan25" <98432@dsprelated> wrote:>>On Tue, 25 Mar 2014 22:17:03 -0500, "Dan25" <98432@dsprelated> wrote: >> >>>Hi, >>> >>>I have a general about the usage of equalization for non-coherent >systems. >>>I'm considering Differential PSK for a single carrier SISO system. I >>>consider different channels with different no., of taps. But I'm not >sure >>>what type of equalizer should be used for a such a system. >>> >>>Thanks. >>> >>>_____________________________ >>>Posted through www.DSPRelated.com >> >>The easiest thing to do is use a Constant Modulus Algorithm. There >>are downsides to CMA, but it works reasonably well at high SNR. >> >>Otherwise it gets very tricky. With some special tricks one can get >>reasonably close to coherent system performance, even at lower SNRs. >> >>This is an implementation of my favorite method: >> >>http://www.anchorhill.com/LinkedFiles/AH1000_MPRE_datasheet_v1.0.pdf >> >> >>Eric Jacobsen >>Anchor Hill Communications >>http://www.anchorhill.com >> > >Hi Eric, > >Thanks very much, Is there any good reference to implement the CMA >algorithm ? Are there any algorithms which operate using Channel State >Information ? > >Best regards.A web search on CMA using relevant terms will probably get you what you need. Also look at the performance tradeoffs involved with CMA, as it does have limitations. If you already have Channel State Information then you don't really need a training or adaptation algorithm, just transform the CSI into filter coefficients. Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
Reply by ●March 26, 20142014-03-26
On Wednesday, March 26, 2014 1:27:28 AM UTC-4, Eric Jacobsen wrote:> > The easiest thing to do is use a Constant Modulus Algorithm. There > are downsides to CMA, but it works reasonably well at high SNR. > > Otherwise it gets very tricky. With some special tricks one can get > reasonably close to coherent system performance, even at lower SNRs. > > This is an implementation of my favorite method: > > http://www.anchorhill.com/LinkedFiles/AH1000_MPRE_datasheet_v1.0.pdf >To the OP's benefit: Even if the bits-to-symbols mapping is differential, the detection of the symbols can still be done coherently. It all depends on the exact setup. When you are working on the symbols agnostic to how the bits are mapped to the symbols, then all you need to care about is how the symbols are relative to the constellation. Nice that Eric cited one of the few Viterbi & Viterbi papers ;-).
