Multicarrier Downlink Transmission for High-speed Railway in Non-Gaussian Noise Channels

We consider a wireless wideband downlink transmission for high-speed railway. In this scenario, the wireless channel is double selective, i.e., multipath and highly time-varying, and involves non-Gaussian noise at the receiver side. Conventional receiver algorithms designed for Gaussian noise and time-invariant channels are inefficient especially for impulsive noise and time-varying multipath channels, which are the main challenges in high-speed railway. They may lead to communication outage and involve large carrier offsets, respectively. Regarding the challenges, a synthesis scheme is proposed to resolve these issues in this paper. First, in order to mitigate the impulsive noise, three impulsive noise detection algorithms are adopted to discriminate impulsive noise from the received signal. The blanking algorithm is used to suppress the discriminated noise. Then, the refined signal goes through a well-designed filtering, which transforms the received signal with varying frequency offsets into an angle domain. The advantage in the angle domain is that the frequency offset in each angle becomes nearly constant. Thus, the frequency offsets can be estimated within each angle to realize accurate estimation. In the simulation, three different impulsive noise models are considered to evaluate the communication reliability in this circumstance. The simulation results show that the proposed scheme significantly improves the reliable performance of the high-speed railway communication systems.