On the net DoF comparison between ZF and MAT over time-varying MISO broadcast channels

We compare two classes of linear precoding strategies, zero-forcing (ZF) beamforming and the “MAT” scheme recently proposed by Maddah-Ali and Tse, over time-varying MISO broadcast channels. For a temporally correlated fading process with band-limited Doppler spectrum in [-F, F], with 0 ≤ F <; 1/2, it has been shown that ZF achieves a fraction 1-2F of the optimal degrees of freedom (DoF). On the other hand, MAT exploits delayed channel state information at transmitter (CSIT) and guarantees a constant DoF irrespectively of the fading Doppler bandwidth. In this work, we compare the net DoFs of both schemes by accounting for the correlation between fading blocks and for the resource required for downlink channel estimation. It is found that the downlink training for each receiver to learn the CSI of the others might be detrimental for the performance of MAT especially when the system size (number of users and transmit antennas) is not much smaller than the channel block length.

[1]  Nihar Jindal MIMO broadcast channels with finite rate feedback , 2005, GLOBECOM.

[2]  Sean A. Ramprashad,et al.  Multi-user MIMO with outdated CSI: Training, feedback and scheduling , 2011, 2011 49th Annual Allerton Conference on Communication, Control, and Computing (Allerton).

[3]  Giuseppe Caire,et al.  Multiuser MIMO Achievable Rates With Downlink Training and Channel State Feedback , 2007, IEEE Transactions on Information Theory.

[4]  S. Shamai,et al.  On the Required Accuracy of Transmitter Channel State Information in Multiple Antenna Broadcast Channels , 2007, 2007 Conference Record of the Forty-First Asilomar Conference on Signals, Systems and Computers.

[5]  Giuseppe Caire,et al.  Training and Feedback Optimization for Multiuser MIMO Downlink , 2009, IEEE Transactions on Communications.

[6]  Jeffrey G. Andrews,et al.  MISO Broadcast Channels with Delayed Finite-Rate Feedback: Predict or Observe? , 2011, IEEE Transactions on Wireless Communications.

[7]  Syed A. Jafar,et al.  Interference Alignment: A New Look at Signal Dimensions in a Communication Network , 2011, Found. Trends Commun. Inf. Theory.

[8]  Mohammad Ali Maddah-Ali,et al.  Completely Stale Transmitter Channel State Information is Still Very Useful , 2010, IEEE Transactions on Information Theory.