In wireless communications and when you have a SISO system, multipath is considered to be an impairment according to theory.
Suppose now that we have an indoor SISO Rician channel (a strong line of sight path accompanied with some multipath), can the existence of these multipaths under some circumstances be of any benefit when compared with a channel that has no multipaths, i.e. only a direct line of site path?
If you can't discriminate amongst the paths you can't use them constructively. If you can't guarantee that they won't interfere with your LOS signal (as is usually the case) then they are generally an impairment.
MIMO essentially uses the resolving power of an antenna array to send differing signals along the different paths which can be received and discriminated by the receiving array in such a way to benefit the system.
There is another way though, and that is temporal discrimination - different paths typically have different propagation delays. Look up RAKE receiver for a CDMA system.
Whether or not you can use this depends on the bandwidth of your channel - if your bandwidth is too small the multipath can give flat fading and nothing you can do will help.
After some comments where temporal discrimination is classified as a type of MIMO, whilst one could argue that this has merit, my understanding is that the usual meaning of MIMO in wireless communications is applied to the use of antenna array(s) to resolve the multipath.
From Wikipedia:
In radio, multiple-input and multiple-output, or MIMO, is a method for multiplying the capacity of a radio link using multiple transmission and receiving antennas to exploit multipath propagation.
multipath is considered to be an impairment according to theory.
No. Theory does not "judge". Some human makes an assessment and calls it an impairment. We're all free to disagree.
can the existence of these multipaths under some circumstances be of any benefit when compared with a channel that has no multipaths, i.e. only a direct line of site path?
Definitely is an advantage, you have a lower outage probability with multiple paths.
However, the moment you start to consider your multiple paths as being somewhat uncorrelated realizations of the channel producing different channel outputs that you try to make use of to get a lower outage probability:
Congratulations! You've just transformed your SISO into a SIMO channel. No matter how much you try to call it SISO, it's no longer single-output :)
So, if you want to call it SISO, you can't make use of the statistical independence of paths, because once you do that, it becomes a SIMO system.
In effect: when you call a channel "multipath SISO", you're constraining yourself to not making use of the independent paths. So, then there's nothing "good" you can make out of this. You can, with the help of an equalizer in the ideal case only become nearly as good as the single-path SISO case.
Once you stop calling it SISO, you can make use of the nice math that -MO brings. This is not a property of the channel, but of the systems we mean when we say SISO or SIMO.
