How about higher sampling rate in adaptive antenna array

Hi,
In smart antenna domain, the sampling rate normally is the same at the
symbol rate. In TDMA systems, there are not many sync symbols which can be
used as reference signals. In order to accormodate to the convergence of the
adaptive algorithm, some longer sync symbol information is required.
In my view, the symbol has been shaped to preserve spectrum efficiency. Why
can't we use more sampling rate in the view of an analog waveform restoring?
In fact, using the same rate of symbol rate is only half of Nyquist sampling
rate at most. If we use higher sampling rate, the benifit we get cannot
compensate for the higher implementation cost? One book mentions the
fractionally spaced equalizers. They have the above ideas. I think adaptive
antennas are similar to equalizers to some point. Can adaptive antennas use
the fractional sampling method? Or, the fractional sampling is just
appropriate for equalizers to dela with the accurate timing problem.

Any suggestion is appreciated.

You can view a multiple (receive) antenna system as a spatial
equalizer.  If you sample your received signal at a higher rate,
i.e. have more sensors (antennas) at the receiver, then the beamformer
resolution is going to improve.  You can also look at it from
the point of view of creating more artificial multipaths in your
system.  So that if the received signal from one antenna (or more)
experience a spectral null, if you have more antennas, then you
increase your chance of correctly receiving the signal because another
antenna may receive a version of the signal that is free from
the effects of the spectral null.  So from a technical point of view,
there is benefit of having more (receive) antennas.  However,
as you mention, you would incur a higher implementation cost.  Also,
if you are talking doing this on a mobile terminal, then this will not
be feasible because of the increase in power usage which will drain
more power from the battery. 

cf 

Jeff wrote:
> 
> Hi,
> In smart antenna domain, the sampling rate normally is the same at the
> symbol rate. In TDMA systems, there are not many sync symbols which can be
> used as reference signals. In order to accormodate to the convergence of the
> adaptive algorithm, some longer sync symbol information is required.
> In my view, the symbol has been shaped to preserve spectrum efficiency. Why
> can't we use more sampling rate in the view of an analog waveform restoring?
> In fact, using the same rate of symbol rate is only half of Nyquist sampling
> rate at most. If we use higher sampling rate, the benifit we get cannot
> compensate for the higher implementation cost? One book mentions the
> fractionally spaced equalizers. They have the above ideas. I think adaptive
> antennas are similar to equalizers to some point. Can adaptive antennas use
> the fractional sampling method? Or, the fractional sampling is just
> appropriate for equalizers to dela with the accurate timing problem.
> 
> Any suggestion is appreciated.