Hi, I am actually new in sparse signal processing and first trying to get the basics of the sparse signals. In OFDM systems the sparse channel is estimated by using some sparse signal processing algorithms such as compressive sensing. Could anyone please explain me how to model the sparse channel/sparse signal in OFDM systems? What is the logic behind the sparse representation? Thanks in advance! --------------------------------------- Posted through http://www.DSPRelated.com
Sparse Channel Estimation in OFDM Systems
Started by ●May 12, 2015
Reply by ●May 12, 20152015-05-12
On Tue, 12 May 2015 02:11:43 -0500, "runinrainy" <100271@DSPRelated> wrote:>Hi, > >I am actually new in sparse signal processing and first trying to get the >basics of the sparse signals. In OFDM systems the sparse channel is >estimated by using some sparse signal processing algorithms >such as compressive sensing. >Could anyone please explain me how to model the sparse channel/sparse >signal in OFDM systems? What is the logic behind the sparse >representation?>Thanks in advance!Usually in DSP "sparse" just means that you don't have a fully-populated set of data for the space it may need to fill. That may or may not mean that something is "undersampled", and often it does not mean that. In the case of a channel matrix for, for example, a MIMO system, it may mean that the matrix is not fully populated but is populated enough to get good results. In a SISO OFDM system sparse channel estimation may mean that you have occassional pilot tones for channel estimation, but that they do not occupy or sample all of the subcarriers so that some interpolation is needed between them. Compressive sensing is another topic that probably has additional points of view on the term "sparse". You probably need to be a little more specific about what you're trying to do. Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
Reply by ●May 14, 20152015-05-14
>On Tue, 12 May 2015 02:11:43 -0500, "runinrainy" <100271@DSPRelated> >wrote: > >>Hi, >> >>I am actually new in sparse signal processing and first trying to get >the >>basics of the sparse signals. In OFDM systems the sparse channel is >>estimated by using some sparse signal processing algorithms >>such as compressive sensing. >>Could anyone please explain me how to model the sparse channel/sparse >>signal in OFDM systems? What is the logic behind the sparse >>representation? > >>Thanks in advance! > >Usually in DSP "sparse" just means that you don't have a >fully-populated set of data for the space it may need to fill. That >may or may not mean that something is "undersampled", and often it >does not mean that. > >In the case of a channel matrix for, for example, a MIMO system, it >may mean that the matrix is not fully populated but is populated >enough to get good results. > >In a SISO OFDM system sparse channel estimation may mean that you have >occassional pilot tones for channel estimation, but that they do not >occupy or sample all of the subcarriers so that some interpolation is >needed between them. > >Compressive sensing is another topic that probably has additional >points of view on the term "sparse". > >You probably need to be a little more specific about what you're >trying to do. > > >Eric Jacobsen >Anchor Hill Communications >http://www.anchorhill.comThank you Eric! As far as I see from my search for the basics of sparse signal representation, for example y=QX where X is the sparse signal that we are trying to estimate it, there is a dictionary matrix which I am confused about and am in difficulty understanding the logic behind it. Appreciate if you could also help me understanding this. Thanks again! --------------------------------------- Posted through http://www.DSPRelated.com
