Measurement-based analysis of cross-layer adaptation for MIMO in mobile urban environments

The performance of mobile multiple-input-multiple-output (MIMO) systems is determined both by the instantaneous channel characteristics and by the rate of change of those characteristics. The local environment and operating frequency affect the spatial channel response, which impacts the ability of the system to sustain data throughput, while the dynamics of the channel response impact the accuracy of the available channel state information, limiting the supportable packet length. Measured MIMO data from two frequencies in the UHF band are characterised and used to investigate the dependence of system performance on the changing spatial channel responses. It is demonstrated that to maximise average throughput, taking into account retransmissions because of packet failures, a trade-off is required among the data rate, training overhead and packet length, which depends on both the operating frequency and the local environment. This work emphasises the importance both of adaptation in MIMO communication systems and of considering realistic channel conditions in their design and evaluation.