[1-2] but it is still a concern when implementing MU-MIMO concepts into existing standards like LTE .
MU-MIMO and carrier aggregation have been widely perceived as the primary means to dramatically improve mobile broadband services and to support wider transmission bandwidths. MU-MIMO is already supported in LTE Release 8 via transmission mode 5 (TM5) (wherein eNB employs two antennas). In theory MU-MIMO provides throughput gains that scale linearly with the number of antennas at the eNB . However, several problems such as the residual multi-user interference seen at the user equipment (UE) need to be addressed  in the case of TM5.
Compared to single-user MIMO, the spatial separation between users is an advantage as the channels of these users tend to be less correlated. However, given the limited codebook size, the inter-user interference experienced by the UEs cannot be jointly processed at the receive side. Thus, no efficient interference mitigation is possible [3-4]. Without adequate interference mitigation, the benefits of MU-MIMO cannot be achieved. As we described in previous blog posts, interference management techniques like zero-forcing precoding  can be carried out by the eNB if the downlink channel estimates are available. Unfortunately, obtaining the required channel state information estimates at the eNB require a substantial amount of feedback bandwidth in the uplink. This led LTE Release 8 to adopt a simple codebook-based precoding scheme in TM5. While it can reduce the inter-user interference, there will still be some amount of interference experienced by the users. Therefore, to achieve the MU-MIMO gains in LTE systems, the UEs need to implement efficient interference cancelation techniques as proposed in  and investigated in the Spectrum Aggregation and Multi-user MIMO: Real-World Impact (SAMURAI) project .
The SAMURAI project focuses on the practical aspects of carrier aggregation and MU-MIMO techniques in LTE and LTE-Advanced systems. One outcome from this project is related to system level gain. It has been shown that a gain of up to 20% in cell throughput is expected when using an interference-aware receiver , at the expense of some loss in cell edge throughput performance. This has also been reported in Figure 11.8 in  as the trade-off between throughput and coverage when using beamforming.
An example on what one would expect from the use of an interference aware receiver is depicted below. In this case, a downlink fast fading channel with the dual-antenna eNodeB and two single-antenna UEs with a 3GPP LTE rate 1/3 turbo code is used with different puncturing patterns. In Figure 1, "IA Rx" and "SU Rx" represent the respective low-complexity interference-aware receiver and the single-user receiver . In this test scenario, the sum rates are fixed (i.e. if two users are served with QPSK with rate 1/2 in the multi-user mode, then one user is served with QAM16 with rate 1/2 in the single-user mode, thereby equating the sum rate in both cases to 2 bps/Hz) .
Figure 1. MU MIMO performance in fading channel with the dual-antenna eNodeB and two single-antenna UEs .
For more details, we refer the reader to [3-4].
 M. Ahmed Ouameur "MU-MIMO part 1"
 M. Ahmed Ouameur "MU-MIMO part 2"
 R Ghaffar and R. Knopp, " Interference-aware receiver structure for multiuser MIMO and LTE," EURASIP Journal on Wireless Communications and Networking 2011, 2011:40.
 B. BADIC, A. F. CATTONI, M. DIEUDONNE, J. DUPLICY, P. FAZEKAS, F. KALTENBERGER, I. Z. KOVÁCS, G. VIVIER, " Advances in Carrier Aggregation and Multi- User MIMO for LTE-Advanced: Outcomes from SAMURAI project," xxx.
 S. Sisia, I. Toufik and M. Baker, "LTE – The UMTS Long Term Evolution: From Theory to Practice," August 29, 2nd edition.