Late last month a guy called
Steve Perlman announced of a new technology called DIDO (Distributed-Input-Distributed-Output) that could revolutionise the way wireless transmission works and can help fix the channel capacity problem as described by Shannon's formula. A whitepaper describing this technology is available
here.
I havent gone through the paper in any detail nor do I understand this DIDO very well but what many experienced engineers have pointed out is that this is MU-MIMO in disguise. Without going into any controversies, lets look at MU-MIMO as its destined to play an important part in LTE-Advanced (the real '4G').
Also, I have been asked time and again about this Shannon's channel capacity formula. This formula is better known by its name
Shannon-Hartley theorem. It states:
C <= B log2 (1 + S/N)
where:
C = channel capacity (bits per second)
B = bandwidth (hertz)
S/N = Signal to Noise ratio (SNR)
In a good channel, SNR will be high. Take for example a case when SNR is 20db then log2 (1 + 100) = 6.6. In an extremely noisy channel SNR will be low which would in turn reduce the channel capacity.
In should be pointed out that the Shannon's formula holds true for all wireless technologies except for when multiuser transmission like MU-MIMO (or DIDO) is used.
Anyway, I gave a
simple explanation on MU-MIMO before. Another simple explanation of what an MU-MIMO is as explained in this video below:
The picture below (
from NTT) gives a good summary of the different kinds of MIMO technology and their advantages and disadvantages. More details could be read from
here.
Click to enlarge
As we can see, MU-MIMO is great but it is complex in implementation.
Click to enlarge
Multiuser MIMO technology makes it possible to raise wireless transmission speed by increasing the number of antennas at the base station, without consuming more frequency bandwidth or increasing modulation multiple-values. It is therefore a promising technology for incorporating broadband wireless transmission that will be seamlessly connected with wired transmission in the micro waveband (currently used for mobile phones and wireless LAN, and well suited to mobile communications use), where frequency resources are in danger of depletion. Since it also allows multiple users to be connected simultaneously, it is seen as a solution to the problem specific to wireless communications, namely, slow or unavailable connections when the number of terminals in the same area increases (see Figure 9 above).
There is a good whitepaper in NTT Docomo technical journal that talks about Precoding and Scheduling techniques for increasing the capacity of MIMO channels. Its available
here. There is also a simple explanation of MIMO including MU-MIMO on RadioElectronics
here. If you want to do a bit more indepth study of MU-MIMO then there is a very good research paper in the EURASIP Journal that is available
here (Click on Full text PDF on right for FREE download).
Finally, there is a
3GPP study item on MIMO Enhancements for LTE-Advanced which is a Release-11 item that will hopefully be completed by next year. That report should give a lot more detail about how practical would it be to implement it as part of LTE-Advanced. The following is the justification of doing this study:
The Rel-8 MIMO and subsequent MIMO enhancements in Rel-10 were designed mostly with homogenous macro deployment in mind. Recently, the need to enhance performance also for non-uniform network deployments (e.g. heterogeneous deployment) has grown. It would therefore be beneficial to study and optimize the MIMO performance for non-uniform deployments where the channel conditions especially for low-power node deployments might typically differ from what is normally encountered in scenarios considered so far.
Downlink MIMO in LTE-Advanced has been enhanced in Release 10 to support 8-layer SU-MIMO transmission and dynamic SU-MU MIMO switching. For the 8-tx antenna case, the CSI feedback to support downlink MIMO has been enhanced with a new dual-codebook structure aimed at improving CSI accuracy at the eNB without increasing the feedback overhead excessively. Precoded reference symbols are provided for data demodulation, allowing arbitrary precoders to be used by the eNB for transmission. In many deployment scenarios, less than 8 tx antennas will be employed. It is important to focus on the eNB antenna configurations of highest priority for network operators.
The enhancement of MIMO performance through improved CSI feedback for high priority scenarios not directly targeted by the feedback enhancements in Release 10, especially the case of 4 tx antennas in a cross-polarised configuration, in both homogeneous and heterogeneous scenarios should be studied.
MU-MIMO operation is considered by many network operators as important to further enhance system capacity. It is therefore worth studying further potential enhancement for MU-MIMO, which includes UE CSI feedback enhancement and control signaling enhancement. Furthermore, open-loop MIMO enhancements were briefly mentioned but not thoroughly investigated in Rel-10.
In addition, the experience from real-life deployments in the field has increased significantly since Rel-8. It would be beneficial to discuss the experience from commercial MIMO deployments, and identify if there are any potential short-comings and possible ways to address those. For example, it can be discussed if robust rank adaptation works properly in practice with current UE procedures that allow a single subframe of data to determine the rank. In addition the impact of calibration error on the performance could be discussed.
This work will allow 3GPP to keep MIMO up to date with latest deployments and experience.
