Does diversity helps channel tracking?

Hello,
Recently I read a paper
http://calypso.unicauca.edu.co/gntt/grupo/maestria/MaterialBandaAnchaInalambrica/channel_models/1385clar.pdf

The result is: with receiver diversity, adaptive RLS frequency-domain
equalization(and channel tracking) can track speed up to 100 miles/hour,
while its signle antenna counterpart has poor tracking capability. 
I wonder why diversity can improve tracking. Furthermore, does diversity
helps time domain channel tracking?
Thanks! 

On Jun 23, 8:07&#4294967295;am, "buaawh" <eewu...@gmail.com> wrote:
> Hello,
> Recently I read a paperhttp://calypso.unicauca.edu.co/gntt/grupo/maestria/MaterialBandaAncha...
>
> The result is: with receiver diversity, adaptive RLS frequency-domain
> equalization(and channel tracking) can track speed up to 100 miles/hour,
> while its signle antenna counterpart has poor tracking capability.
> I wonder why diversity can improve tracking. Furthermore, does diversity
> helps time domain channel tracking?
> Thanks!

Diversity techniques are tricky to evaluate fairly because a lot of
times
it depends on how exactly the problem setup is defined.  So I did not
read the linked paper, but rather I'll propose the following
observation.

Any tracking / locationing problem involves some sort of inversion at
some point, for example in converting the time-of-arrival or angle-of-
arrival estimates into a 3-D location.  It is known that such problems
depend on the relative geometries of the sensors and the targets.

So sure, with more sensors you improve the distribution of the
conditioning number required for the inversion.  But how much
better you can do is dependent on the problem setup.