Before selecting the best antennas in MIMO, the channels from all antennas are measured/estimated. Definitely if the number of antennas increase with comparatively less number of RF chains, the switching between all antenna combinations increases.
So is there any method to decrease this effect by selecting best antennas and simultaneously estimating channel statistics?
Antenna subset selection without CSI is somewhat possible for receiver antenna selection. Look the well known publications below:
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=1341263
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=1237409
It may be possible to combine blind channel estimation techniques with the antenna selection algorithms to achieve the same at the transmitter side. But I am not an expert in that.
Having said that, you can always find new ways where you can pose the problem in a different way. In essence, this can be posed as an exploration vs exploitation problem, where you can explore the subset combinations and also exploit the best subsets possible. The definition of best can change over time. So, a balance has to be struck between the two phases. What you achieve by doing this is that instead of exhaustive search across all possible subsets, you reduce your search by sequential exploration and reduce the amount of channel training required for CSI acquisition. The two papers highlight similar ideas.
http://www.eurasip.org/Proceedings/Eusipco/Eusipco2012/Conference/papers/1569585537.pdf
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6574205
- Badges
- Science topic