In order to enable advanced downlink multiantenna techniques, like MU-MIMO or Network MIMO, accurate channel knowledge at the transmitter is required. In order to save scarce uplink resources, the feedback has to be compressed. This paper provides a practical channel feedback compression technique which exploits the sparse nature of radio channels and removes redundant and irrelevant information. It introduces a novel path detection algorithm by modifying the existing orthogonal matching pursuit algorithm. Furthermore it presents efficient practical bit loading techniques. Channel aging effects can be efficiently handled by prediction, as shown in this paper. The prediction integrates very well into the presented feedback compression framework. For application to existing radio propagation channel compression, measurement results are discussed, dealing with the question how sparse the actual channel is.
展开▼
