Forums

Equalizers implemented in GSM receivers

Started by Dan25 March 18, 2014
Hi All,

I need some specific information on channel equalizer implementations of
practical GSM receivers. I assume time domain equalizers are used in GSM
phones, but I would like to the equalizer algorithms and training
algorithms and convergence algorithms used by them. 

I highly appreciate if anyone can share his/her experience on this regard.
All the references are highly appreciated.

Thanks in advance.	 

_____________________________		
Posted through www.DSPRelated.com
On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated> wrote:

>Hi All, > >I need some specific information on channel equalizer implementations of >practical GSM receivers. I assume time domain equalizers are used in GSM >phones, but I would like to the equalizer algorithms and training >algorithms and convergence algorithms used by them. > >I highly appreciate if anyone can share his/her experience on this regard. >All the references are highly appreciated. > >Thanks in advance.
Receiver details for specific impementations are usually kept proprietary, so you may not be able to get much info on what is actually done, and it probably varies a fair amount from implementation to implementation. There are training sequences built into the GSM frame structure, so there are a lot of different tecqniques that can be used to train the equalizer. Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
>On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated> wrote: > >>Hi All, >> >>I need some specific information on channel equalizer implementations of >>practical GSM receivers. I assume time domain equalizers are used in GSM >>phones, but I would like to the equalizer algorithms and training >>algorithms and convergence algorithms used by them. >> >>I highly appreciate if anyone can share his/her experience on this
regard.
>>All the references are highly appreciated. >> >>Thanks in advance. > >Receiver details for specific impementations are usually kept >proprietary, so you may not be able to get much info on what is >actually done, and it probably varies a fair amount from >implementation to implementation. > >There are training sequences built into the GSM frame structure, so >there are a lot of different tecqniques that can be used to train the >equalizer. > >Eric Jacobsen >Anchor Hill Communications >http://www.anchorhill.com >
I'm pretty sure that the intended "equalization" method for a GSM burst is maximum likelihood sequence estimation. -Doug _____________________________ Posted through www.DSPRelated.com
On Tue, 18 Mar 2014 13:33:59 -0500, "DougB" <60916@dsprelated> wrote:

>>On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated> wrote: >> >>>Hi All, >>> >>>I need some specific information on channel equalizer implementations of >>>practical GSM receivers. I assume time domain equalizers are used in GSM >>>phones, but I would like to the equalizer algorithms and training >>>algorithms and convergence algorithms used by them. >>> >>>I highly appreciate if anyone can share his/her experience on this >regard. >>>All the references are highly appreciated. >>> >>>Thanks in advance. >> >>Receiver details for specific impementations are usually kept >>proprietary, so you may not be able to get much info on what is >>actually done, and it probably varies a fair amount from >>implementation to implementation. >> >>There are training sequences built into the GSM frame structure, so >>there are a lot of different tecqniques that can be used to train the >>equalizer. >> >>Eric Jacobsen >>Anchor Hill Communications >>http://www.anchorhill.com >> >I'm pretty sure that the intended "equalization" method for a GSM burst is >maximum likelihood sequence estimation. > >-Doug
The standards don't specify implementation, just the waveform. Implementers are free to do whatever they wish in the receiver. Sometimes the air interface waveform is designed with some candidate implementation in mind, but one is always free to do their own tricks in the receiver.
>_____________________________ >Posted through www.DSPRelated.com
Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
>On Tue, 18 Mar 2014 13:33:59 -0500, "DougB" <60916@dsprelated> wrote: > >>>On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated> wrote: >>> >>>>Hi All, >>>> >>>>I need some specific information on channel equalizer implementations
of
>>>>practical GSM receivers. I assume time domain equalizers are used in
GSM
>>>>phones, but I would like to the equalizer algorithms and training >>>>algorithms and convergence algorithms used by them. >>>> >>>>I highly appreciate if anyone can share his/her experience on this >>regard. >>>>All the references are highly appreciated. >>>> >>>>Thanks in advance. >>> >>>Receiver details for specific impementations are usually kept >>>proprietary, so you may not be able to get much info on what is >>>actually done, and it probably varies a fair amount from >>>implementation to implementation. >>> >>>There are training sequences built into the GSM frame structure, so >>>there are a lot of different tecqniques that can be used to train the >>>equalizer. >>> >>>Eric Jacobsen >>>Anchor Hill Communications >>>http://www.anchorhill.com >>> >>I'm pretty sure that the intended "equalization" method for a GSM burst
is
>>maximum likelihood sequence estimation. >> >>-Doug > >The standards don't specify implementation, just the waveform. >Implementers are free to do whatever they wish in the receiver. >Sometimes the air interface waveform is designed with some candidate >implementation in mind, but one is always free to do their own tricks >in the receiver. > >>_____________________________ >>Posted through www.DSPRelated.com > >Eric Jacobsen >Anchor Hill Communications >http://www.anchorhill.com >
Thanks Doug and Eric. Do you have any references which I can use to find the implementations ? I want to get an accurate number for complexity of channel equalizer and estimator algorithms. _____________________________ Posted through www.DSPRelated.com
On Tuesday, March 18, 2014 9:06:05 PM UTC-4, Dan25 wrote:
> >On Tue, 18 Mar 2014 13:33:59 -0500, "DougB" <60916@dsprelated> wrote: > > > > > >>>On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated> wrote: > > >>> > > >>>>Hi All, > > >>>> > > >>>>I need some specific information on channel equalizer implementations > > of > > >>>>practical GSM receivers. I assume time domain equalizers are used in > > GSM > > >>>>phones, but I would like to the equalizer algorithms and training > > >>>>algorithms and convergence algorithms used by them. > > >>>> > > >>>>I highly appreciate if anyone can share his/her experience on this > > >>regard. > > >>>>All the references are highly appreciated. > > >>>> > > >>>>Thanks in advance. > > >>> > > >>>Receiver details for specific impementations are usually kept > > >>>proprietary, so you may not be able to get much info on what is > > >>>actually done, and it probably varies a fair amount from > > >>>implementation to implementation. > > >>> > > >>>There are training sequences built into the GSM frame structure, so > > >>>there are a lot of different tecqniques that can be used to train the > > >>>equalizer. > > >>> > > >>>Eric Jacobsen > > >>>Anchor Hill Communications > > >>>http://www.anchorhill.com > > >>> > > >>I'm pretty sure that the intended "equalization" method for a GSM burst > > is > > >>maximum likelihood sequence estimation. > > >> > > >>-Doug > > > > > >The standards don't specify implementation, just the waveform. > > >Implementers are free to do whatever they wish in the receiver. > > >Sometimes the air interface waveform is designed with some candidate > > >implementation in mind, but one is always free to do their own tricks > > >in the receiver. > > > > > >>_____________________________ > > >>Posted through www.DSPRelated.com > > > > > >Eric Jacobsen > > >Anchor Hill Communications > > >http://www.anchorhill.com > > > > > > > Thanks Doug and Eric. > > > > Do you have any references which I can use to find the implementations ? I > > want to get an accurate number for complexity of channel equalizer and > > estimator algorithms. > > > > _____________________________ > > Posted through www.DSPRelated.com
Although as others have pointed out GSM standards don't specify receiver implementation, the waveform is tailored for MLSE. Additionally, from first hand experience I know that GSM's equipment conformance specification pushes the theoretical performance bounds of a 16 state MSLE. So while manufacturers may choose their implementations, it's a tall order for suboptimal designs. The following is a good reference in that it contains a thorough treatment of Viterbi equalization for phase modulation in general and GMSK in particular. The book is out of print but if you can get your hands on a used copy I think you'll get a lot of mileage out of it. Raymond Steele, Mobile Radio Communications, IEEE Press, 1992, ISBN 0-7803-1102-7. Some other references: J.C.S. Cheung, R. Steele, "Modified Viterbi equalizer for mobile radio channels having large multi-path delay", Electronics Letters, vol. 25, no. 19, pp. 1309-1311, 14 Sept, 1989. N.S. Hoult, C.A. Dace, A.P. Cheer, "Implementation of an equaliser for the GSM system", Proc. of the 5th Int. Conf. on Radio Receivers Associated Systems, Cambridge, U.K., 24-26 July 1990. R.D'Avella, L. Moreno, M. Sant'Agostino, "An adaptive MLSE receiver for TDMA digital mobile radio", IEEE Journal on Selected Areas in Communications, vol. 7, no. 1, pp. 122-129, Jan, 1989.
>On Tuesday, March 18, 2014 9:06:05 PM UTC-4, Dan25 wrote: >> >On Tue, 18 Mar 2014 13:33:59 -0500, "DougB" <60916@dsprelated> wrote: >>=20 >> > >>=20 >> >>>On Tue, 18 Mar 2014 02:51:09 -0500, "Dan25" <98432@dsprelated>
wrote:
>>=20 >> >>> >>=20 >> >>>>Hi All, >>=20 >> >>>> >>=20 >> >>>>I need some specific information on channel equalizer
implementations
>>=20 >> of >>=20 >> >>>>practical GSM receivers. I assume time domain equalizers are used
in
>>=20 >> GSM >>=20 >> >>>>phones, but I would like to the equalizer algorithms and training >>=20 >> >>>>algorithms and convergence algorithms used by them.=20 >>=20 >> >>>> >>=20 >> >>>>I highly appreciate if anyone can share his/her experience on this >>=20 >> >>regard. >>=20 >> >>>>All the references are highly appreciated. >>=20 >> >>>> >>=20 >> >>>>Thanks in advance. =20 >>=20 >> >>> >>=20 >> >>>Receiver details for specific impementations are usually kept >>=20 >> >>>proprietary, so you may not be able to get much info on what is >>=20 >> >>>actually done, and it probably varies a fair amount from >>=20 >> >>>implementation to implementation. >>=20 >> >>> >>=20 >> >>>There are training sequences built into the GSM frame structure, so >>=20 >> >>>there are a lot of different tecqniques that can be used to train
the
>>=20 >> >>>equalizer. >>=20 >> >>> >>=20 >> >>>Eric Jacobsen >>=20 >> >>>Anchor Hill Communications >>=20 >> >>>http://www.anchorhill.com >>=20 >> >>> >>=20 >> >>I'm pretty sure that the intended "equalization" method for a GSM
burst
>>=20 >> is >>=20 >> >>maximum likelihood sequence estimation. >>=20 >> >> >>=20 >> >>-Doug =20 >>=20 >> > >>=20 >> >The standards don't specify implementation, just the waveform. >>=20 >> >Implementers are free to do whatever they wish in the receiver. >>=20 >> >Sometimes the air interface waveform is designed with some candidate >>=20 >> >implementation in mind, but one is always free to do their own tricks >>=20 >> >in the receiver. >>=20 >> > >>=20 >> >>_____________________________ =09 >>=20 >> >>Posted through www.DSPRelated.com >>=20 >> > >>=20 >> >Eric Jacobsen >>=20 >> >Anchor Hill Communications >>=20 >> >http://www.anchorhill.com >>=20 >> > >>=20 >>=20 >>=20 >> Thanks Doug and Eric. >>=20 >>=20 >>=20 >> Do you have any references which I can use to find the implementations ?
=
>I >>=20 >> want to get an accurate number for complexity of channel equalizer and >>=20 >> estimator algorithms. =20 >>=20 >>=20 >>=20 >> _____________________________ =09 >>=20 >> Posted through www.DSPRelated.com > >Although as others have pointed out GSM standards don't specify receiver
im=
>plementation, the waveform is tailored for MLSE. Additionally, from first
=
>hand experience I know that GSM's equipment conformance specification
pushe=
>s the theoretical performance bounds of a 16 state MSLE. So while
manufact=
>urers may choose their implementations, it's a tall order for suboptimal
de=
>signs. > >The following is a good reference in that it contains a thorough treatment
=
>of Viterbi equalization for phase modulation in general and GMSK in
particu=
>lar. The book is out of print but if you can get your hands on a used
copy=
> I think you'll get a lot of mileage out of it. > >Raymond Steele, Mobile Radio Communications, IEEE Press, 1992, ISBN
0-7803-=
>1102-7. > >Some other references: > >J.C.S. Cheung, R. Steele, "Modified Viterbi equalizer for mobile radio
chan=
>nels having large multi-path delay", Electronics Letters, vol. 25, no. 19,
=
>pp. 1309-1311, 14 Sept, 1989. > >N.S. Hoult, C.A. Dace, A.P. Cheer, "Implementation of an equaliser for the
=
>GSM system", Proc. of the 5th Int. Conf. on Radio Receivers Associated
Syst=
>ems, Cambridge, U.K., 24-26 July 1990. > >R.D'Avella, L. Moreno, M. Sant'Agostino, "An adaptive MLSE receiver for
TDM=
>A digital mobile radio", IEEE Journal on Selected Areas in Communications,
=
>vol. 7, no. 1, pp. 122-129, Jan, 1989. > >
Thanks very much. Its very useful. _____________________________ Posted through www.DSPRelated.com