Robust adaptive modulation with imperfect channel information

Adaptive modulation is a promising technique to increase system throughput considerably. However, it relies on perfect channel state information (CSI), and is sensitive to errors in CSI. In this work, we maximize the system transmission rate based on a lower bound of average bit error rate (BER) while satisfying the transmit power and BER constraint. In order to further enhance the system throughput, adaptive modulation scheme is combined with a robust transmit beamformer to obtain extra diversity gain. Moreover, to pay the penalty for the lower bound of the average BER, we introduce a probabilistic constraint by keeping a low outage probability of signal-to-noise ratio (SNR). Simulation results show that the proposed scheme provides the maximum system throughput compared with several state-of-the-art robust adaptive schemes, and always guarantees the target BER.