TA-HAN: Towards RIS-Aided Cell-Free Distributed Time Adjustment for Asynchronous Transceiver

Cell-free is seen as one of the most important technology for future wireless communications, and it can be enhanced by reconfigurable intelligent surface (RIS) to provide better coverage and service. However, the distributed deployment of access points (APs) and RISs bring heterogeneous propagation delays, which cannot be ignored in low-latency communications considering the possible inter-symbol interference and inter-carrier interference. Although there are several previous works about robust precoding against time offset, they can only compensate the performance loss when the delay differences are not large. What is more important, per-sub-carrier compensation is not affordable in wideband. In this correspondence, we equip each AP/RIS with a single delay to adjust the transceiving timing directly, and propose a heterogeneous graph neural network, called time adjustment heterogeneous graph attention network (TA-HAN), to optimize the delay values. Our TA-HAN can be deployed in a distributed style to support distributed time adjustment, and fits dynamic setups with uniformly near-optimal performance.