Feedback Reduction for Spatial Multiplexing with Linear Precoding

This paper presents two novel methods to optimally compress the feedback for spatial multiplexing with linear precoding. The methods exploit the time correlation of the channel and the knowledge of the previously fed back precoder matrices to estimate the conditional probabilities of the different possible feedback indices. These probabilities are then used to losslessly compress the actual feedback using variable-length codes. Two compression schemes are presented, one for a non-dedicated feedback channel and one for a dedicated feedback channel.