Robust JPEG2000 image transmission over closed-loop MIMO-OFDM with limited feedback

In traditional MIMO-OFDM beamforming system with limited feedback, the channel state information (CSI) feedback from the receiver contains only the selected beamforming vectors information. In this paper, we propose to include the SNR strength knowledge of some subcarriers in the feedback and implement adaptive channel selection (ACS) at the transmitter for JPEG2000 image transmission. As a result, a significant enhancement and robustness can be achieved. We also propose nonuniform-based beamforming algorithms to further reduce the number of required feedback bits. Although the simulation results show that this leads to worse bit error rate (BER) performance, the received image quality measured in average peak signal-to-noise ratio (PSNR), however, is comparable with the traditional optimally quantized beamforming which requires much more feedback bits. Moreover, with the proposed algorithms, the possibility of successful image reconstruction at the receiver is much higher than the existing algorithms even at low SNRs.