Cross-layer optimization for a multiuser MIMO audio transmission

In this contribution we investigate the cross-layer issues occurring in a multi-user MIMO scenario. The scenario is based on a geometric model, thus containing realistic correlations in space, time, and frequency. As typical service we assume an audio transmission, where we evaluate the transmission quality by using the PEAQ (perceptual evaluation of audio quality) parameter. To build the bridge from the audio quality to the physical layer we derive a mapping from the PEAQ parameter to operation points consisting of required data rates and maximum allowable frame error rates. The optimization goal is to assure that each user attains a certain quality while minimizing the costs, i.e., bandwidth or transmission power. The parameters to be optimized are the multiple access scheme (SDMA, TDMA), the MIMO transmission mode (beamforming on one or two eigen-beams with diversity combining or multiple data streams, i.e., spatial multiplexing) and the choice of the modulation and coding scheme