부산대학교 도서관

High-speed rail (HSR) travel has become a popular transportation mode for people, and its train-ground communication has received wide attention. However, current research lacks the optimization of communication pilot design in the HSR scenario. To overcome this challenge, this letter establishes a system model for train-ground predictive antenna (PA) communication. We take the sounding reference signal (SRS) design in the uplink channel pilot as a research object, to maximize the system downlink throughput. We decompose the original problem into three subproblems involving the frequency-domain resource allocation for PA, pattern design, and time slot period of SRS to eliminate coupling. Then, we propose the block coordinate descent (BCD)-based coalitional game and a heuristic algorithm, to obtain a reasonable SRS design of PAs so as improve the accuracy of channel estimation for PAs and channel prediction for main antennas. Finally, the effectiveness of the proposed algorithm is verified by simulation and measured data, with superior performance in enhancing the downlink system throughput.