Channel Prediction for Real-Time Wireless Communication with MmWave SC-FDE in IIoT Systems

With the application of wireless sensor-actuator networks in the Industrial Internet of Things (IIoT), it is crucially important to ensure the real-timing of data transmission. The millimeter wave (mmWave) communicating at the extremely high frequency band is a promising solution for the rapidly expanding data throughput in IIoT, due to the wide usable frequency band. In extremely high frequency band, the channel coherent time will be obviously reduced and becomes shorter than the frame duration. In this case, the channel state information (CSI) acquisition based on channel estimation will provide outdated information for coherent signal detection. Therefore, forecasting the channel variation for real-time data transmission is necessary. In this paper, we investigate the channel prediction methods in both the frequency and time domains for mmWave single-carrier frequency-domain-equalization (SC-FDE) systems. In the frequency domain, the channel prediction is conducted on each subcarrier, while the time domain predictor on each channel tap. As a number of the channel taps in the time domain are mainly composed of estimation noise, we separate these channel taps composed of estimation noise from the significant taps before building the prediction model. In this paper, the autoregressive (AR) model is employed to perform the channel prediction in the both domains. The simulation results show that the time domain predictor increases the prediction accuracy while reducing the computation complexity.