Adaptive Downlink OFDMA System With Low-Overhead and Limited Feedback in Time-Varying Underwater Acoustic Channel

Because of the time-varying characteristic and the large propagation delay in underwater acoustic (UWA) channel, adaptive resource allocation in UWA orthogonal frequency-division multiplexing access (OFDMA) system cannot be performed with the assumption of perfect channel state information (CSI) at the transmitter. Therefore, this paper focuses on the outdated CSI and proposes an adaptive downlink OFDMA system with low-overhead and limited feedback in time-varying UWA channel. First, a data fitting method is proposed for channel reconstruction, which shows better performance and lower overhead than the traditional group quantization method in simulations. Second, we define the per-subcarrier channel temporal correlation (PSCTC) and long-term statistical mean value as two indicators of outdated CSI, for optimizing CSI feedback in UWA channel with large propagation delay. Finally, a CSI selection method based on channel correlation attenuation (CCA) factor is proposed, where CSI used for feedback is selected between outdated-instantaneous CSI and outdated-average CSI according to the CCA factor in the current observation period. Sea trail results show that the CSI reconstruction based on data fitting method has lower bit error ratio and less overhead than traditional group quantization method. The PSCTC coefficient is related to the feedback delay, and the long-term statistics is stationary, which means that the outdated-average CSI is suitable for channel feedback. The CSI selection method based on CCA factor has better performance than that only uses outdated-instantaneous CSI or outdated-average CSI under different time-varying channel conditions in experimental results.