Abstract:In this paper, we propose a cooperative time synchronization (CTS) scheme for cell-free massive multiple-input multiple-output (MIMO) systems, where synchronization packets are jointly broadcast by a set of coordinated access points (APs). Thanks to multiple timestamps available at each target node, the proposed CTS scheme enables a much higher accuracy in clock parameters estimation than the conventional non-cooperative synchronization approaches. In addition to the clock parameters between the target nodes and their associated APs, the clock deviations among APs are also jointly estimated, where the clock deviations can be fed back to their associated APs via acknowledgment for further clock adjustment. Furthermore, to mitigate the impact on synchronization performance under the packet loss scenario, a matrix completion-based CTS (MC-CTS) algorithm is proposed that complements the lost timestamp information. Simulation results demonstrate that the proposed CTS scheme presents a robust performance against the clock deviations among coordinated APs with packet overhead of approximately 50% less than that of the non-broadcast synchronization approaches, and the proposed MC-CTS algorithm effectively enhances the estimation performance of clock parameters, compared to just utilizing the CTS algorithm. Hence, the proposed algorithms are particularly suitable for time-sensitive cell-free massive MIMO systems.

Cooperative Time Synchronization and Robust Clock Parameters Estimation for Time-Sensitive Cell-Free Massive MIMO Systems